CN114579750A - Information processing method and device, computer equipment and storage medium - Google Patents
Information processing method and device, computer equipment and storage mediumDownload PDFInfo
- Publication number
- CN114579750A CN114579750ACN202210314713.XACN202210314713ACN114579750ACN 114579750 ACN114579750 ACN 114579750ACN 202210314713 ACN202210314713 ACN 202210314713ACN 114579750 ACN114579750 ACN 114579750A
- Authority
- CN
- China
- Prior art keywords
- appeal
- target
- processing
- scene
- emotion
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
- G06F16/353—Clustering; Classification into predefined classes
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/334—Query execution
- G06F16/3344—Query execution using natural language analysis
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- General Health & Medical Sciences (AREA)
- Machine Translation (AREA)
Abstract
Description
Translated fromChinese技术领域technical field
本申请涉及数据分析技术领域，尤其涉及一种信息处理方法、装置、计算机设备及存储介质。The present application relates to the technical field of data analysis, and in particular, to an information processing method, apparatus, computer equipment and storage medium.
背景技术Background technique
传统的诉求处理流程复杂，比如金融公司、银行等的客服人员工作繁重复杂，其中一个重要原因是：需要了解各种业务场景，根据客户的咨询或者投诉等，来判断客户询问的具体事项，根据具体的问题归属，将问题再转给不同属主以进行处理，然而，发明人发现，这个过程需要耗费大量的人力、时间，并且由于人员的判断能力参差不齐，很容易一个环节出问题，拖慢整个诉求流程的处理进度，影响用户的体验，甚至影响公司整体形象。因此，迫切需要一套智能化的方案提升诉求处理效率以及诉求质量。The traditional appeal processing process is complicated. For example, the customer service staff of financial companies and banks have heavy and complicated work. One of the important reasons is that they need to understand various business scenarios, and judge the specific matters inquired by customers based on customer consultation or complaints. However, the inventor found that this process requires a lot of manpower and time, and due to the uneven judgment ability of personnel, it is easy to cause problems in one link. Slow down the processing progress of the entire appeal process, affect the user experience, and even affect the overall image of the company. Therefore, there is an urgent need for a set of intelligent solutions to improve the efficiency of appeal processing and the quality of appeals.
发明内容SUMMARY OF THE INVENTION
本申请实施例提供了一种信息处理方法、装置、计算机设备及存储介质，可以提升诉求处理效率和诉求处理质量。The embodiments of the present application provide an information processing method, apparatus, computer equipment, and storage medium, which can improve appeal processing efficiency and appeal processing quality.
一方面，本申请实施例提供了一种信息处理方法，包括：On the one hand, an embodiment of the present application provides an information processing method, including:
获取目标对象的诉求文本，并对所述诉求文本进行分词处理，得到分词结果；Obtain the appeal text of the target object, and perform word segmentation processing on the appeal text to obtain a word segmentation result;
对所述分词结果进行情感识别，得到所述目标对象的目标诉求情感信息，并对所述分词结果进行意图识别，得到所述目标对象的目标诉求意图；Perform emotion recognition on the word segmentation result to obtain the target appeal emotional information of the target object, and perform intention recognition on the word segmentation result to obtain the target appeal intention of the target object;
根据所述目标诉求情感信息和所述目标诉求意图确定目标诉求场景，并根据所述目标诉求场景从第一语料库中匹配出目标诉求处理方案，所述第一语料库包括多种诉求处理方案，每种诉求处理方案对应一种或多种诉求场景；A target appeal scene is determined according to the target appeal emotional information and the target appeal intention, and a target appeal processing plan is matched from a first corpus according to the target appeal scene. The first corpus includes a variety of appeal processing plans, each Each appeal processing scheme corresponds to one or more appeal scenarios;
按照所述目标诉求处理方案进行诉求处理。The appeal is processed according to the target appeal processing plan.
再一方面，本申请实施例提供了一种信息处理装置，包括：In another aspect, an embodiment of the present application provides an information processing apparatus, including:
获取模块，用于获取目标对象的诉求文本，并对所述诉求文本进行分词处理，得到分词结果；an acquisition module, used for acquiring the appeal text of the target object, and performing word segmentation processing on the appeal text to obtain a word segmentation result;
识别模块，用于对所述分词结果进行情感识别，得到所述目标对象的目标诉求情感信息，并对所述分词结果进行意图识别，得到所述目标对象的目标诉求意图；an identification module, configured to perform emotion recognition on the word segmentation result, obtain the target appeal emotion information of the target object, and perform intention recognition on the word segmentation result to obtain the target appeal intention of the target object;
处理模块，用于根据所述目标诉求情感信息和所述目标诉求意图确定目标诉求场景，并根据所述目标诉求场景从第一语料库中匹配出目标诉求处理方案，所述第一语料库包括多种诉求处理方案，每种诉求处理方案对应一种或多种诉求场景；A processing module, configured to determine a target appeal scene according to the target appeal emotional information and the target appeal intention, and match a target appeal processing plan from a first corpus according to the target appeal scene, where the first corpus includes a variety of Appeal handling plan, each appeal handling plan corresponds to one or more appeal scenarios;
处理模块，还用于按照所述目标诉求处理方案进行诉求处理。The processing module is further configured to process appeals according to the target appeal processing scheme.
再一方面，本申请实施例提供了一种计算机设备，包括处理器和存储器，所述处理器和所述存储器相互连接，其中，所述存储器用于存储计算机程序指令，所述处理器被配置用于执行所述程序指令，实现所述的信息处理方法。In another aspect, an embodiment of the present application provides a computer device, including a processor and a memory, wherein the processor and the memory are connected to each other, wherein the memory is used to store computer program instructions, and the processor is configured It is used for executing the program instructions to realize the information processing method.
再一方面，本申请实施例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中存储有计算机程序指令，所述计算机程序指令被处理器执行时，用于执行所述的信息处理方法。In another aspect, an embodiment of the present application provides a computer-readable storage medium, where computer program instructions are stored in the computer-readable storage medium, and when the computer program instructions are executed by a processor, are used to execute the information Approach.
综上所述，计算机设备可以对目标对象的诉求文本进行分词处理，得到分词结果，然后对分词结果进行情感识别，得到目标对象的目标诉求情感信息，并对分词结果进行意图识别，得到目标对象的目标诉求意图，之后可以根据目标诉求情感信息和目标诉求意图确定目标诉求场景，并根据目标诉求场景从第一语料库中匹配出目标诉求处理方案，进而按照目标诉求处理方案进行诉求处理，该过程通过上述智能化诉求处理的流程可以提高诉求处理效率和诉求处理质量。To sum up, the computer equipment can perform word segmentation processing on the appeal text of the target object, obtain the word segmentation result, and then perform emotion recognition on the word segmentation result to obtain the target appeal emotional information of the target object, and perform intention recognition on the word segmentation result to obtain the target object. After that, the target appeal scene can be determined according to the target appeal emotional information and the target appeal intention, and the target appeal processing plan can be matched from the first corpus according to the target appeal scene, and then the appeal processing is carried out according to the target appeal processing plan. Through the above-mentioned intelligent appeal processing process, the appeal processing efficiency and appeal processing quality can be improved.
附图说明Description of drawings
为了更清楚地说明本申请实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本申请的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only These are some embodiments of the present application. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings without any creative effort.
图1是本申请实施例提供的一种信息处理过程的示意图；1 is a schematic diagram of an information processing process provided by an embodiment of the present application;
图2是本申请实施例提供的一种信息处理方法的流程示意图；2 is a schematic flowchart of an information processing method provided by an embodiment of the present application;
图3是本申请再一实施例提供的一种信息处理方法的流程示意图；3 is a schematic flowchart of an information processing method provided by still another embodiment of the present application;
图4是本申请实施例提供的一种信息处理装置的结构示意图；4 is a schematic structural diagram of an information processing apparatus provided by an embodiment of the present application;
图5是本申请实施例提供的一种计算机设备的结构示意图。FIG. 5 is a schematic structural diagram of a computer device provided by an embodiment of the present application.
具体实施方式Detailed ways
下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行描述。The technical solutions in the embodiments of the present application will be described below with reference to the accompanying drawings in the embodiments of the present application.
本申请实施例提供了一种信息处理方案，主要依赖于AI人工智能技术，打破了传统针对对象的诉求(如用户的诉求)进行处理的方式，真正地将问题的判断、处理，交给了计算机来处理，在解放人力的同时，大大提高了对诉求的处理速度。其中，所述信息处理方案可以由计算机设备来执行，所述的计算机设备可以为智能终端或服务器。智能终端可以为台式电脑等具备信息处理能力的智能终端。服务器可以是独立的服务器，也可以是多个物理服务器构成的服务器集群或者分布式系统，还可以是提供云服务、云数据库、云计算、云函数、云存储、网络服务、云通信、中间件服务、域名服务、安全服务、内容分发网络(ContentDelivery Network，CDN)、以及大数据和人工智能平台等基础云计算服务的云服务器，但并不局限于此。The embodiment of the present application provides an information processing solution, which mainly relies on AI artificial intelligence technology, breaks the traditional method of processing the appeal of the object (such as the appeal of the user), and truly transfers the judgment and processing of the problem to the hands of Computer processing, while liberating manpower, greatly improves the processing speed of demands. Wherein, the information processing solution may be executed by a computer device, and the computer device may be an intelligent terminal or a server. The intelligent terminal may be an intelligent terminal with information processing capability such as a desktop computer. The server can be an independent server, or a server cluster or distributed system composed of multiple physical servers, or it can provide cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware Cloud servers for basic cloud computing services such as services, domain name services, security services, Content Delivery Network (CDN), and big data and artificial intelligence platforms, but not limited to this.
下面以对象为用户为例对所述的信息处理方案进行简单的介绍。The information processing scheme described below is briefly introduced by taking the object as the user as an example.
如图1所示，所述的信息处理方案主要包含三大模块：数据识别、场景分析、方案落地。具体地：数据识别模块用于为场景分析模块提供最原始的数据；场景分析模块用于为根据这些原始数据做分析计算，得到计算结果；方案落地模块用于提供该计算结果对应的解决方案，其中，方案落地模块存储了各种各样的解决方案，是问题处理方式的集合。As shown in Figure 1, the described information processing scheme mainly includes three modules: data identification, scene analysis, and scheme implementation. Specifically: the data identification module is used to provide the most original data for the scene analysis module; the scene analysis module is used to perform analysis and calculation based on these original data to obtain the calculation result; the solution implementation module is used to provide the solution corresponding to the calculation result, Among them, the solution landing module stores various solutions, which is a collection of problem processing methods.
其中，所述的数据识别模块包含情感分析、意图识别等功能。在一个实施例中。数据识别模块还包含了语音识别功能。本申请实施例中，可以获取采集到的用户的诉求语音，通过语音识别功能可以将诉求语音转化为诉求文本。其中，诉求文本可以用于表示用户的诉求。除了通过上述方案获取到诉求文本，也可以通过其它方式获取到诉求文本，本申请实施例对获取诉求文本的方式不做限制。通过数据识别模块处理得到的数据，如情感信息为后续模块的处理提供了数据基础。Wherein, the data recognition module includes functions such as sentiment analysis and intention recognition. In one embodiment. The data recognition module also includes a speech recognition function. In the embodiment of the present application, the collected appeal voice of the user can be acquired, and the appeal voice can be converted into appeal text through the speech recognition function. Among them, the appeal text can be used to express the user's appeal. In addition to obtaining the appeal text through the above solution, the appeal text may also be obtained in other ways, and the embodiment of the present application does not limit the method of obtaining the appeal text. The data processed by the data recognition module, such as emotional information, provides the data basis for the processing of the subsequent modules.
其中，所述的场景分析模块包括语料喂养与场景分类等功能。为了更加精准的处理用户诉求，首先要对不同用户的诉求进行分类，细化到具体的场景，例如，通过对不同用户的诉求进行分析，最终细化出的场景可以包括用户咨询产品详情的场景和用户投诉资金延迟到账的场景等场景。其中，场景分析的准确性依赖两种数据，一种是数据识别模块处理得到的数据，另外一种是系统内部的语料喂养。基于数据识别模块处理得到的数据可以用于确定具体的诉求场景。语料喂养可以是指，将运营人员、开发人员、测试人员等在公司运作过程中所有与用户交互的场景，翻译成计算机理解的语言以添加至知识库。随着喂养语料增多，知识库将会越来越完备，从而能更加精准的服务于场景分析。The scene analysis module includes functions such as corpus feeding and scene classification. In order to handle user demands more accurately, the demands of different users must first be classified and refined to specific scenarios. For example, by analyzing the demands of different users, the final refined scenarios can include scenarios where users consult product details. Scenarios such as users complaining about the delayed arrival of funds. Among them, the accuracy of scene analysis depends on two kinds of data, one is the data processed by the data recognition module, and the other is the corpus feeding inside the system. The data processed by the data identification module can be used to determine specific appeal scenarios. Corpus feeding can refer to translating all the scenarios that operators, developers, testers, etc. interact with users during the operation of the company into a language understood by computers to be added to the knowledge base. As the feeding corpus increases, the knowledge base will become more and more complete, so that it can serve the scene analysis more accurately.
其中，所述的方案落地模块致力于提供对诉求的处理方案。每一个与用户交互的场景的解决方案并不是一成不变的，例如咨询类场景的初始解决方案，可能仅仅将用户咨询的产品信息告诉用户，随着知识库的丰富，还可以提供给用户相同类型的其它产品信息等等。对于用户的每一次提问，系统可以计算出至少三种解决方案供运营人员选择，从而匹配出适合不同的用户人群的解决方案。Among them, the solution implementation module is dedicated to providing a solution to the appeal. The solution for each scenario that interacts with users is not static. For example, the initial solution for consulting scenarios may only tell the user the product information consulted by the user. With the enrichment of the knowledge base, it can also provide users with the same type of solutions. Other product information, etc. For each user's question, the system can calculate at least three solutions for operators to choose, so as to match solutions suitable for different user groups.
采用本申请实施例提供的信息处理方案，可以极大提高对诉求的处理速度，并提高问题处理的准确度。相比于传统的纯人工处理，加入了机器学习的优势，随着处理诉求的数量的增长，准确率会越来越高。Using the information processing solution provided by the embodiments of the present application can greatly improve the processing speed of requests and improve the accuracy of problem processing. Compared with the traditional pure manual processing, the advantages of machine learning are added. As the number of processing demands increases, the accuracy rate will become higher and higher.
基于上述描述的信息处理方案，本申请实施例提供了一种信息处理方法。该信息处理方法可以应用于前述提及的计算机设备。参见图2，该方法可以包括以下步骤：Based on the information processing solution described above, an embodiment of the present application provides an information processing method. This information processing method can be applied to the aforementioned computer device. Referring to Figure 2, the method may include the following steps:
S201、获取目标对象的诉求文本，并对所述诉求文本进行分词处理，得到分词结果。S201. Obtain the appeal text of the target object, and perform word segmentation processing on the appeal text to obtain a word segmentation result.
其中，对象可以是用户。相应地，目标对象可以为目标用户。目标用户可以指提交了诉求信息的用户，诉求信息可以为诉求文本或诉求语音。Among them, the object can be a user. Accordingly, the target object may be the target user. The target user may refer to the user who submitted the appeal information, and the appeal information may be the appeal text or the appeal voice.
在一个实施例中，诉求文本可以是目标用户通过第一窗口提交的指定类型的文本。第一窗口可以是聊天窗口、留言窗口或短信窗口。指定类型包括但不限于咨询类、投诉类、表扬类等诉求类型。在一个实施例中，诉求文本还可以是对目标用户的诉求语音进行语音识别后得到的指定类型的文本。In one embodiment, the appeal text may be a specified type of text submitted by the target user through the first window. The first window may be a chat window, a message window or a short message window. Specified types include but are not limited to consultation, complaint, praise and other appeal types. In one embodiment, the appeal text may also be a specified type of text obtained by performing speech recognition on the appeal speech of the target user.
S202、对所述分词结果进行情感识别，得到所述目标对象的目标诉求情感信息，并对所述分词结果进行意图识别，得到所述目标对象的目标诉求意图。S202. Perform emotion recognition on the word segmentation result to obtain target appeal emotion information of the target object, and perform intention recognition on the word segmentation result to obtain the target appeal intention of the target object.
其中，目标诉求情感信息指目标对象的诉求情感信息。目标诉求意图指目标对象的诉求意图。The target appeal emotion information refers to the appeal emotion information of the target object. The target appeal intention refers to the appeal intention of the target object.
在一个实施例中，计算机设备对分词结果进行情感识别，得到目标对象的目标诉求情感信息的方式可以为：计算机设备将分词结果中各个词语与第二语料库中各个词语进行匹配，得到分词结果包括的反映诉求情感的关键词组，然后根据词语与情感熵值的对应关系确定关键词组中各个关键词对应的情感熵值，从而根据各个关键词对应的情感熵值确定目标对象的目标诉求情感信息。关键词组可以包括一个或多个反映诉求情感的关键词。情感熵值可以用于指示情感倾向程度。示例性的，情感熵值可以取0-1之间的数值，情感熵值越高表明肯定意向越强或者态度越积极正面，情感熵值越低表明否定意向越强或者态度越消极负面。由此看出，情感熵值可以用于标识情感状态。In one embodiment, the computer device performs emotion recognition on the word segmentation result, and the method for obtaining the target appeal emotion information of the target object may be: the computer device matches each word in the word segmentation result with each word in the second corpus, and the obtained word segmentation result includes Then, according to the corresponding relationship between words and emotional entropy values, the emotional entropy value corresponding to each keyword in the keyword group is determined, so as to determine the target appeal emotional information of the target object according to the emotional entropy value corresponding to each keyword. The keyword group may include one or more keywords reflecting the sentiment of the appeal. The emotional entropy value can be used to indicate the degree of emotional tendency. Exemplarily, the sentiment entropy value may take a value between 0 and 1. A higher sentiment entropy value indicates a stronger positive intention or a more positive attitude, and a lower sentiment entropy value indicates a stronger negative intention or a more negative attitude. It can be seen that the emotional entropy value can be used to identify the emotional state.
在一个实施例中，计算机设备在关键词组包括一个关键词时，根据所述各个关键词对应的情感熵值确定所述目标对象的目标诉求情感信息的方式可以为：将这个关键词对应的情感熵值确定为目标对象的目标诉求情感信息。在一个实施例中，计算机设备在关键词组包括多个关键词时，根据所述各个关键词对应的情感熵值确定所述目标对象的目标诉求情感信息的方式可以为：计算机设备根据所述各个关键词对应的情感熵值计算得到情感熵值均值，并将所述情感熵值均值确定为所述目标对象的目标诉求情感信息。通过均值计算得到的情感熵值均值能够反映整体的情感倾向程度。举例来说，假设用户甲向计算机设备提交了诉求文本，该诉求文本为“我不喜欢这款产品，这款产品收益低”，计算机设备可以根据此诉求文本获取到关键词组：“不喜欢”“收益低”。这里的“不喜欢”“收益低”属于否定意向的词语。之后，计算机设备可以根据关键词与情感熵值的对应关系，确定这两个关键词分别对应的情感熵值，并对这两个关键词分别对应的情感熵值进行均值计算，得到情感熵值均值以确定为用户甲的诉求情感信息。In one embodiment, when the keyword group includes a keyword, the computer device determines the target appeal emotional information of the target object according to the emotional entropy value corresponding to each keyword. The method may be: The entropy value is determined as the target appeal emotional information of the target object. In one embodiment, when the keyword group includes a plurality of keywords, the computer device determines the target appeal emotional information of the target object according to the emotional entropy value corresponding to each keyword. The method may be: The sentiment entropy value corresponding to the keyword is calculated to obtain the mean sentiment entropy value, and the mean sentiment entropy value is determined as the target appeal sentiment information of the target object. The average emotional entropy value calculated by the average value can reflect the overall emotional tendency. For example, assuming that user A submits an appeal text to the computer device, the appeal text is "I don't like this product, this product has low revenue", the computer device can obtain the keyword group according to the appeal text: "I don't like it" "Low Yield". "Dislike" and "low income" here are words with negative intentions. After that, the computer device can determine the emotional entropy values corresponding to the two keywords according to the corresponding relationship between the keywords and the emotional entropy values, and perform mean calculation on the emotional entropy values corresponding to the two keywords to obtain the emotional entropy value. The mean value is determined as the emotional information of user A's appeal.
在一个实施例中，由于第二语料库记录的关键词可能不完善，因此还可以通过如下方式确定出遗漏的反映诉求情感的关键词，从而结合前述得到的关键词组确定目标对象的目标诉求情感信息。具体方式如下：计算机设备可以判断分词结果中除关键词组外的各词语中是否存在第一词语，第一词语在第二语料库或第三语料库中存在关联的第二词语。计算机设备在确定分词结果中除关键词组外的各词语中存在第一词语后，可以根据词语与情感熵值的对应关系确定第二词语对应的情感熵值，并根据第二词语对应的情感熵值确定第一词语的情感熵值。计算机设备在得到关键词组对应的情感熵值以及第一词语对应的情感熵值后，计算机设备可以根据关键词组对应的情感熵值和第一词语对应的情感熵值确定目标对象对应的目标情感信息。In one embodiment, since the keywords recorded in the second corpus may not be perfect, the missing keywords reflecting the sentiment of the appeal can also be determined in the following manner, so as to determine the sentiment information of the target object in combination with the keyword group obtained above . The specific method is as follows: the computer device can determine whether the first word exists in each word except the keyword group in the word segmentation result, and the first word has an associated second word in the second corpus or the third corpus. After determining that the first word exists in each word except the keyword group in the word segmentation result, the computer device can determine the emotional entropy value corresponding to the second word according to the corresponding relationship between the word and the emotional entropy value, and determine the emotional entropy value corresponding to the second word according to the emotional entropy value corresponding to the second word. value determines the sentiment entropy value of the first word. After the computer device obtains the emotional entropy value corresponding to the keyword group and the emotional entropy value corresponding to the first word, the computer device can determine the target emotional information corresponding to the target object according to the emotional entropy value corresponding to the keyword group and the emotional entropy value corresponding to the first word. .
在一个实施例中，计算机设备根据第二词语对应的情感熵值获得第一词语对应的情感熵值的方式为计算机设备将第一词语对应的情感熵值设置为与第二词语对应的情感熵值相近的数值(如设置为与第二词语对应的情感熵值相近的数值相差预设数值的一个数值)，这就表明第一词语对应的情感熵值与第二词语对应的情感熵值之间的差值较小。其中，第二词语可以包括第一词语中的词语，或第二词语中的部分词语与第一词语中的部分词语相同。第二词语与第一词语具有相同的情感倾向。举例来说，假设用户甲向计算机设备提交了诉求文本，该诉求文本为“我不耐这款产品，这款产品收益低”，计算机设备可以根据此诉求文本获取到关键词组：“收益低”。计算机设备可以确定分词结果中除“收益低”外的各词语中存在第一词语“不耐”，“不耐”在第二语料库或第三语料库存在关联的第二词“不”，“不耐”和“不”都具有否定倾向。计算机设备可以将“不耐”的情感熵值设置为与“不”对应的情感熵值相接近的数值。In one embodiment, the method for the computer device to obtain the emotional entropy value corresponding to the first word according to the emotional entropy value corresponding to the second word is that the computer device sets the emotional entropy value corresponding to the first word to the emotional entropy corresponding to the second word A value with a similar value (for example, it is set to a value similar to the emotional entropy value corresponding to the second word that is different from a preset value), which indicates that the emotional entropy value corresponding to the first word is equal to the emotional entropy value corresponding to the second word. The difference between is small. Wherein, the second words may include words in the first words, or some words in the second words are the same as some words in the first words. The second word has the same emotional tendency as the first word. For example, suppose user A submits an appeal text to a computer device, and the appeal text is "I'm not patient with this product, this product has low revenue", the computer device can obtain the keyword group according to the appeal text: "low revenue" . The computer device can determine that the first word "impatient" exists in each word except "low income" in the word segmentation result, and the second corpus or the third corpus associated with "impatient" exists in the second word "no", "no". Both "resist" and "no" have negative tendencies. The computer device may set the emotional entropy value of "impatient" to a value close to the emotional entropy value corresponding to "no".
在一个实施例中，计算机设备根据分词结果进行意图识别，得到目标对象的目标诉求意图的方式可以为：计算机设备将分词结果输入意图识别模型以进行意图识别，得到目标对象的目标诉求意图。其中，意图识别模型可以是利用多个诉求文本样本中每个诉求文本样本对应的分词结果对初始的分类模型训练得到的。在一个实施例中，计算机设备根据分词结果进行意图识别，得到目标对象的目标诉求意图的方式还可以为：计算机设备将分词结果以及目标诉求情感信息输入意图识别模型以进行意图识别，得到目标对象的目标诉求意图。In one embodiment, the computer device performs intent recognition according to the word segmentation result to obtain the target appeal intent of the target object: the computer device inputs the word segmentation result into the intent recognition model for intent recognition, and obtains the target object appeal intent. The intent recognition model may be obtained by training the initial classification model by using the word segmentation result corresponding to each appeal text sample in the multiple appeal text samples. In one embodiment, the computer device performs intention recognition according to the word segmentation result, and the method of obtaining the target appeal intention of the target object may also be as follows: the computer device inputs the word segmentation result and the target appeal emotional information into the intention recognition model for intention recognition, and obtains the target object. target appeal intent.
在一个实施例中，若通过意图识别模型无法识别出目标对象的目标诉求意图，则可以将分词结果设置为目标类别，之后对目标类别下的所有分词结果进行聚类处理，得到多个分词结果集合；计算机设备发送处理指示信息至目标智能终端，处理指示信息携带多个分词结果集合，处理指示信息用于指示相关人员，如运营人员针对每个分词结果集合标记对应的诉求意图；计算机设备接收目标智能终端返回的对每个分词结果集合标记的诉求意图，并确定根据目标对象的诉求文本得到的分词结果所在的目标分词结果集合，从而将对目标分词结果集合标记的诉求意图确定为目标对象的目标诉求意图。需要说明的是，目标类别下的所有分词结果中可能聚合有不同来源的分词结果，采用上面这种方式在意图识别模型无法识别对象意图的情况下也能够获得对象的意图，并且在聚类后提供给相关人员，如运营人员进行标记，也能够减轻运营人员的工作量。In one embodiment, if the target appeal intention of the target object cannot be identified through the intention recognition model, the word segmentation result can be set as the target category, and then all the word segmentation results under the target category are clustered to obtain multiple word segmentation results Collection; the computer equipment sends processing instruction information to the target intelligent terminal, the processing instruction information carries multiple word segmentation result sets, and the processing instruction information is used to indicate relevant personnel, such as operators marking the corresponding appeal intent for each word segmentation result set; the computer equipment receives The target intelligent terminal returns the appeal intent for each word segmentation result set mark, and determines the target word segmentation result set where the word segmentation result obtained according to the appeal text of the target object is located, so as to determine the appeal intent marked on the target word segmentation result set as the target object target appeal intent. It should be noted that all word segmentation results under the target category may be aggregated with word segmentation results from different sources. Using the above method, the intent of the object can be obtained even if the intent recognition model cannot recognize the intent of the object, and after clustering, the intent of the object can be obtained. Providing relevant personnel, such as operators for marking, can also reduce the workload of operators.
S203、根据所述目标诉求情感信息和所述目标诉求意图确定目标诉求场景，并根据所述目标诉求场景从第一语料库中匹配出目标诉求处理方案，所述第一语料库包括多种诉求处理方案，每种诉求处理方案对应一种或多种诉求场景。S203. Determine a target appeal scene according to the target appeal emotional information and the target appeal intention, and match a target appeal processing scheme from a first corpus according to the target appeal scene, where the first corpus includes multiple appeal processing schemes , and each appeal processing scheme corresponds to one or more appeal scenarios.
在一个实施例中，第一语料库中的诉求处理方案可以包括以下至少一项：①根据运营人员通过电话等方式记录的诉求内容以及对应的处理方案获得的诉求处理方案；②根据公司开发人员和/或测试人员在处理系统问题时记录的问题原因及处理方案获得的诉求处理方案；③通过网络爬虫获取的诉求场景与对应的处理方案获得的诉求处理方案。In one embodiment, the appeal processing scheme in the first corpus may include at least one of the following: ① according to the appeal content recorded by the operator through telephone and other means and the appeal processing scheme obtained from the corresponding processing scheme; ② according to the company developer and / or the cause of the problem recorded by the tester when dealing with the system problem and the appeal solution obtained by the solution; ③ the appeal scenario obtained through the web crawler and the corresponding solution.
S204、按照所述目标诉求处理方案进行诉求处理。S204, processing the appeal according to the target appeal processing scheme.
本申请实施例中，计算机设备可以确定目标诉求意图关联的至少一个诉求场景，并根据至少一个诉求场景中每个诉求场景对应的诉求情感信息，从至少一个诉求场景中确定出目标情感信息对应的诉求场景以作为目标诉求场景。之后，计算机可以从第一语料库中查询出目标诉求场景对应的目标诉求处理方案，并按照目标诉求处理方案进行诉求处理。举例来说，假设目标诉求场景为查询。计算机设备通过上述过程可以从第一语料库中查询出该目标诉求场景对应的目标诉求处理方案，目标诉求处理方案为向用户提供产品详情，则计算机设备便可以提供产品详情给目标用户。In the embodiment of the present application, the computer device may determine at least one appeal scene associated with the target appeal intent, and determine the target emotional information corresponding to the at least one appeal scene according to the appeal emotion information corresponding to each appeal scene in the at least one appeal scene. The appeal scene is used as the target appeal scene. After that, the computer can query the target appeal processing scheme corresponding to the target appeal scene from the first corpus, and process the appeal according to the target appeal processing scheme. For example, suppose the target appeal scenario is a query. Through the above process, the computer device can query the target appeal processing scheme corresponding to the target appeal scene from the first corpus. The target appeal processing scheme is to provide product details to the user, and then the computer device can provide product details to the target user.
在一个实施例中，所述的至少一个诉求场景可以通过以下方式确定：根据诉求意图与诉求场景的对应关系确定目标诉求意图对应的至少一个诉求场景，以作为目标诉求意图关联的至少一个诉求场景；或，确定目标诉求意图对应的目标诉求类别，并确定目标诉求类别下的至少一个诉求场景，以作为目标诉求意图关联的至少一个诉求场景。示例性的，目标诉求意图为查询产品、投诉产品、表扬产品或表达对产品的期望时，计算机设备可以确定目标诉求意图对应的目标诉求类别为产品，并可以确定产品这个诉求类别下的至少一个诉求场景。在一个实施例中，在确定目标诉求意图对应的至少一个诉求场景的过程中，为了准确的确定出至少一个诉求场景，可以根据诉求意图与诉求场景的对应关系或确定目标诉求类别下的诉求场景集合，然后根据分词结果从诉求场景集合中确定出至少一个诉求场景。In one embodiment, the at least one appeal scene may be determined in the following manner: at least one appeal scene corresponding to the target appeal intent is determined according to the correspondence between the appeal intent and the appeal scene, as at least one appeal scene associated with the target appeal intent ; or, determine the target appeal category corresponding to the target appeal intent, and determine at least one appeal scenario under the target appeal category as at least one appeal scenario associated with the target appeal intent. Exemplarily, when the target appeal intention is to inquire about a product, complain about a product, praise a product, or express an expectation for a product, the computer device may determine that the target appeal category corresponding to the target appeal intention is a product, and may determine at least one product under this appeal category. request scene. In one embodiment, in the process of determining at least one appeal scene corresponding to the target appeal intent, in order to accurately determine the at least one appeal scene, the appeal scene under the target appeal category may be determined according to the correspondence between the appeal intent and the appeal scene set, and then determine at least one appeal scene from the appeal scene set according to the word segmentation result.
在一个实施例中，计算机设备在根据目标诉求方案进行诉求处理的过程中，除了可以自动化智能化进行诉求处理之外，还可以在此之前将目标诉求处理方案发送至目标智能终端，并当接收到目标智能终端根据目标诉求处理方案返回的确认指令时，执行按照目标诉求处理方案对目标对象进行诉求处理的操作。In one embodiment, in the process of processing the appeal according to the target appeal scheme, the computer device can not only automatically and intelligently process the appeal, but also send the target appeal processing scheme to the target intelligent terminal before receiving it. When the confirmation instruction returned by the target intelligent terminal according to the target appeal processing scheme is reached, the operation of processing the appeal of the target object according to the target appeal processing scheme is performed.
在一个实施例中，计算机设备可以将分词结果发送至目标智能终端，计算机设备还可以接收目标智能终端反馈的在分词结果中未识别出的语料，并将该语料更新至前述提及的第二语料库中。In one embodiment, the computer device can send the word segmentation result to the target intelligent terminal, and the computer device can also receive the corpus that is not recognized in the word segmentation result fed back by the target intelligent terminal, and update the corpus to the second mentioned above. in the corpus.
可见，图2所示的实施例中，计算机设备可以对目标对象的诉求文本进行分词处理，得到分词结果，然后对分词结果进行情感识别，得到目标对象的目标诉求情感信息，并对分词结果进行意图识别，得到目标对象的目标诉求意图，之后可以根据目标诉求情感信息和目标诉求意图确定目标诉求场景，并根据目标诉求场景从第一语料库中匹配出目标诉求处理方案，进而按照目标诉求处理方案进行诉求处理，该过程可以提高诉求处理效率和诉求处理质量。It can be seen that in the embodiment shown in FIG. 2, the computer device can perform word segmentation on the appeal text of the target object, obtain a word segmentation result, and then perform emotion recognition on the word segmentation result to obtain the target object's target appeal emotional information, and perform a word segmentation result on the word segmentation result. Intent recognition, obtain the target appeal intention of the target object, and then determine the target appeal scene according to the target appeal emotional information and target appeal intention, and match the target appeal processing plan from the first corpus according to the target appeal scene, and then according to the target appeal processing plan Carry out appeal processing, which can improve the efficiency and quality of appeal processing.
基于上述提供的信息处理方法，本申请实施例还提供了一种信息处理方法。该信息处理方法可以应用于前述提及的计算机设备。参见图3，该方法可以包括以下步骤：Based on the information processing method provided above, an embodiment of the present application further provides an information processing method. This information processing method can be applied to the aforementioned computer device. Referring to Figure 3, the method may include the following steps:
S301、获取目标对象的诉求文本，并对所述诉求文本进行分词处理，得到分词结果。S301. Obtain the appeal text of the target object, and perform word segmentation processing on the appeal text to obtain a word segmentation result.
S302、对所述分词结果进行情感识别，得到所述目标对象的目标诉求情感信息，并对所述分词结果进行意图识别，得到所述目标对象的目标诉求意图。S302. Perform emotion recognition on the word segmentation result to obtain target appeal emotion information of the target object, and perform intention recognition on the word segmentation result to obtain the target appeal intention of the target object.
S303、根据所述目标诉求情感信息和所述目标诉求意图确定目标诉求场景。S303. Determine a target appeal scene according to the target appeal emotion information and the target appeal intention.
其中，步骤S301-步骤S303可以参见图2实施例中的步骤S201-步骤S203，本申请实施例在此不做赘述。Wherein, for steps S301 to S303, reference may be made to steps S201 to S203 in the embodiment of FIG. 2 , and details are not described herein in this embodiment of the present application.
S304、获取所述目标对象的对象数据。S304. Acquire object data of the target object.
在对象为用户时，对象数据可以为用户数据。需要说明的是，本申请所涉及的用户数据，均为经用户授权或者经过各方充分授权的信息和数据，且相关数据的收集、使用和处理需要遵守相关国家和地区的相关法律法规和标准。When the object is a user, the object data may be user data. It should be noted that the user data involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data need to comply with the relevant laws, regulations and standards of the relevant countries and regions. .
其中，所述的用户数据可以为性别、年龄、行为偏好等数据。The user data may be data such as gender, age, and behavior preference.
S305、根据所述对象数据提取对象标签。S305. Extract object tags according to the object data.
在对象为用户时，对象标签可以为用户标签。When the object is a user, the object label can be a user label.
在一个实施例中，在用户数据包括性别时，用户标签可以包括性别标签，在用户数据包括年龄时，用户标签可以包括年龄段标签(如70后，80后，90后或老年、中年、青年)，在用户数据包括行为偏好时，用户标签可以包括行为偏好标签。In one embodiment, when the user data includes gender, the user label may include a gender label, and when the user data includes age, the user label may include an age group label (such as post-70s, post-80s, post-90s or old, middle-aged, youth), where user data includes behavioral preferences, user tags may include behavioral preference tags.
S306、从所述至少一种诉求处理方案中筛选出与所述对象标签匹配的诉求处理方案；S306. Screen out an appeal processing scheme matching the object label from the at least one appeal processing scheme;
S307、将所述匹配的诉求处理方案确定为目标诉求处理方案。S307. Determine the matched appeal processing scheme as the target appeal processing scheme.
在一个实施例中，诉求处理方案为向用户进行信息反馈，当用户标签包括性别标签时，不同性别标签对应的诉求处理方案对应提供的内容详细程度不同和/或内容类型不同。当用户标签包括年龄段标签时，不同年龄段标签对应的诉求处理方案对应提供的内容详细程度不同和/或内容类型不同。当用户标签包括行为偏好标签时，不同行为偏好标签对应提供的诉求处理方案的内容详细程度不同和/或内容类型不同。In one embodiment, the appeal processing solution is to provide information feedback to the user, and when the user tag includes a gender tag, the appeal processing solutions corresponding to different gender tags provide different content details and/or different content types. When the user tag includes an age group tag, the appeal processing solutions corresponding to different age group tags provide different levels of content detail and/or different content types. When the user tag includes a behavior preference tag, the content details and/or content types of the appeal processing solutions provided corresponding to different behavior preference tags are different.
举例来说，当用户标签包括老年时，可以从目标诉求场景对应的诉求处理方案中识别出与老年标签匹配的诉求处理方案，假设至少一种诉求处理方案为多种诉求处理方案时，每种目标诉求处理方案向用户提供的产品详情不同，而此处识别出的老年标签匹配的诉求处理方案可以是多种诉求处理方案中产品详情最为详细的诉求处理方案。For example, when the user tag includes the elderly, the appeal processing scheme that matches the elderly tag can be identified from the appeal processing schemes corresponding to the target appeal scene. It is assumed that at least one appeal processing scheme is a variety of appeal processing schemes. The target appeal processing solutions provide users with different product details, and the appeal processing solution identified here for the elderly label matching can be the appeal processing solution with the most detailed product details among the various appeal processing solutions.
在一个实施例中，诉求处理方案为向用户提供电子资源，如电子优惠券、补贴，当用户标签包括性别标签时，不同性别标签对应的诉求处理方案对应提供的电子资源的数值或电子资源的类型不同。当用户标签包括年龄段标签时，不同年龄段标签对应提供的诉求处理方案的电子资源的数值或电子资源的类型不同。当用户标签包括行为偏好标签时，不同行为偏好标签对应提供的诉求处理方案电子资源的数值或电子资源的类型不同。In one embodiment, the appeal processing solution is to provide users with electronic resources, such as electronic coupons and subsidies. When the user tags include gender tags, the appeal processing solutions corresponding to different gender tags correspond to the value of the provided electronic resources or the value of the electronic resources provided. different types. When the user tag includes an age group tag, different age group tags provide different values or types of electronic resources corresponding to the request processing solution. When the user tag includes a behavior preference tag, the values or types of electronic resources provided corresponding to different behavior preference tags are different.
S308、按照所述目标诉求处理方案进行诉求处理。S308, processing the appeal according to the target appeal processing scheme.
其中，步骤S308可以参见图2实施例中的步骤S204，本申请实施例在此不做赘述。Wherein, for step S308, reference may be made to step S204 in the embodiment of FIG. 2 , which is not repeated in this embodiment of the present application.
在一个实施例中，除了可以通过提取用户标签来匹配诉求处理方案，计算机设备还可以从分词结果中提取产品标签，根据产品标签从至少一种处理方案中确定出目标处理方案。在一个实施例中，计算机设备可以根据产品标签确定产品类型，然后从第一语料库中查询出目标诉求场景对应的至少一种诉求处理方案，并从至少一种诉求处理方案中筛选出与产品类型匹配的诉求处理方案，从而将匹配的诉求处理方案确定为目标诉求处理方案。举例来说，当产品标签为已下架时，计算机设备可以从至少一种诉求处理方案中识别出与下架匹配的诉求处理方案，该匹配的诉求处理方案可以为向用户推荐新的产品，或可以为向用户推荐新的同类型的产品。In one embodiment, in addition to extracting user tags to match appeal processing schemes, the computer device can also extract product tags from word segmentation results, and determine a target processing scheme from at least one processing scheme according to the product tags. In one embodiment, the computer device may determine the product type according to the product label, then query the first corpus to find at least one appeal processing scheme corresponding to the target appeal scenario, and filter out the product type from the at least one appeal processing scheme. The matching appeal processing scheme is determined, so that the matching appeal processing scheme is determined as the target appeal processing scheme. For example, when the product is labeled as being off the shelf, the computer device can identify a request processing solution matching the removal from at least one request processing solution, and the matching request processing solution can be recommending a new product to the user, Or it can recommend new products of the same type to users.
在一个实施例中，除了可以获取目标对象的文本，还可以获取目标对象的诉求图像，诉求图像可以为表情图片等图像。计算机设备还可以对诉求图像进行情感识别，得到诉求图像的诉求情感信息。相应地，目标诉求场景除了可以根据目标情感信息和目标诉求意图确定，还可以通过如下方式确定：计算机设备可以根据诉求图像的诉求情感信息、目标诉求情感信息和目标诉求意图确定目标诉求场景。在一个实施例中，计算机设备利用诉求图像的诉求情感信息更新目标情感信息，得到更新后的诉求情感信息。根据更新后的诉求情感信息和目标诉求意图确定目标诉求场景。其中，根据更新后的诉求情感信息和目标诉求意图确定目标诉求场景的方式可以参见根据目标诉求情感信息和目标诉求意图确定目标诉求场景的方式，本申请实施例在此不做赘述。In one embodiment, in addition to acquiring the text of the target object, an appeal image of the target object may also be acquired, and the appeal image may be an image such as an expression picture. The computer device can also perform emotion recognition on the appeal image to obtain appeal emotion information of the appeal image. Correspondingly, in addition to determining the target appeal scene according to the target emotional information and the target appeal intention, the computer device can also determine the target appeal scene according to the appeal emotional information of the appeal image, the target appeal emotional information and the target appeal intention. In one embodiment, the computer device uses the appeal emotion information of the appeal image to update the target emotion information to obtain the updated appeal emotion information. The target appeal scene is determined according to the updated appeal emotion information and the target appeal intention. The method of determining the target appeal scene according to the updated appeal emotion information and the target appeal intention may refer to the method of determining the target appeal scene according to the target appeal emotional information and the target appeal intention, which is not repeated in this embodiment of the present application.
在一个实施例中，计算机设备可以通过情感信息识别模型对诉求图像进行情感识别，得到诉求图像的诉求情感信息。其中，所述的情感信息识别模型是利用多个诉求图像样本对初始的回归模型进行训练后得到的。在一个实施例中，计算机设备还可以通过情感信息识别模型对诉求情感图像进行情感识别，得到诉求图像的初始诉求情感信息。考虑到在实际的应用场景中单一依靠诉求情感模型识别出诉求图像的诉求情感信息可能不够准确，比如，对于微笑这个表情图片，一些情况下这个表情图片表示正面积极的情绪，在另一些情况下这个词则表情负面消极的情绪，因此计算机设备还可以将分词结果或输入情绪类型识别模型，得到情绪类别识别模型识别出的诉求情感类型，之后计算机设备利用识别出的诉求情感类型对诉求图像的初始诉求情感信息进行调整，得到诉求图像对应的调整后的诉求情感信息，以作为诉求图像对应的诉求情感信息。比如，在根据情绪类别识别模型识别出的诉求情感类型确定诉求图像表示的为正面积极的情绪，则可以将初始诉求情感信息调大一定数值，反之，则可以将初始诉求情感信息调小一定数值。In one embodiment, the computer device may perform emotion recognition on the appeal image through an emotion information recognition model to obtain appeal emotion information of the appeal image. Wherein, the emotion information recognition model is obtained after training the initial regression model by using a plurality of appeal image samples. In one embodiment, the computer device may also perform emotion recognition on the appeal emotion image through the emotion information recognition model to obtain initial appeal emotion information of the appeal image. Considering that in actual application scenarios, relying solely on the appeal emotion model to identify the appeal emotion information of the appeal image may not be accurate enough, for example, for the smiley expression picture, in some cases this expression picture represents positive emotions, in other cases This word expresses negative emotions. Therefore, the computer device can also input the word segmentation result or input the emotion type recognition model to obtain the appeal emotion type identified by the emotion type recognition model. The initial appeal emotion information is adjusted to obtain the adjusted appeal emotion information corresponding to the appeal image, which is used as the appeal emotion information corresponding to the appeal image. For example, when it is determined that the appeal image represents positive emotions according to the appeal emotion type identified by the emotion category recognition model, the initial appeal emotional information can be increased by a certain value, otherwise, the initial appeal emotional information can be reduced by a certain value .
在一个实施例中，前述提及的对象还可以是项目或产品。诉求文本还可以为用于表达对目标项目或目标产品的诉求的文本。该诉求文本可以是由目标人员通过第二窗口(可以用于编辑诉求文本的窗口)提交的或还可以是对目标人员的诉求语音进行语音识别后的得到的。所述的对象数据可以为项目数据或产品数据。所述的诉求意图可以是具体的开发意图。所述的诉求场景可以是具体的开发场景。所述的诉求处理方案可以是具体的开发方案。In one embodiment, the aforementioned objects may also be items or products. The appeal text may also be a text used to express an appeal to the target item or target product. The appeal text may be submitted by the target person through the second window (a window that can be used to edit the appeal text) or may also be obtained by performing speech recognition on the appeal voice of the target person. The object data can be item data or product data. The appeal intent may be a specific development intent. The appeal scenario may be a specific development scenario. The claim processing solution may be a specific development solution.
可见，图3所示的实施例中，计算机设备还可以结合用户标签来确定具体的诉求处理方案，从而根据确定出的诉求处理方案进行更为精准的诉求处理，进而提升诉求处理质量。It can be seen that in the embodiment shown in FIG. 3 , the computer device can also determine a specific appeal processing scheme in combination with the user tag, so as to perform more accurate appeal processing according to the determined appeal processing scheme, thereby improving the appeal processing quality.
本申请涉及区块链技术，如所述的诉求文本或诉求图像可以是从区块链读取的，或者通过区块链网络获取到的。This application relates to blockchain technology, and the appeal text or appeal image as described may be read from the blockchain or obtained through the blockchain network.
基于前述描述的信息处理方法，本申请实施例还提供了一种信息处理装置。该信息处理装置可以应用于前述提及的计算机设备。具体地，参见图4，该装置可以包括：Based on the information processing method described above, an embodiment of the present application further provides an information processing apparatus. The information processing apparatus can be applied to the aforementioned computer equipment. Specifically, referring to FIG. 4, the apparatus may include:
获取模块401，用于获取目标对象的诉求文本，并对所述诉求文本进行分词处理，得到分词结果。The obtaining
识别模块402，用于对所述分词结果进行情感识别，得到所述目标对象的目标诉求情感信息，并对所述分词结果进行意图识别，得到所述目标对象的目标诉求意图。The
处理模块403，用于根据所述目标诉求情感信息和所述目标诉求意图确定目标诉求场景，并根据所述目标诉求场景从第一语料库中匹配出目标诉求处理方案，所述第一语料库包括多种诉求处理方案，每种诉求处理方案对应一种或多种诉求场景。The
处理模块403，还用于按照所述目标诉求处理方案进行诉求处理。The
在一种可选的实施方式中，识别模块402对所述分词结果进行情感识别，得到所述目标对象的目标诉求情感信息，具体为将所述分词结果中各个词语与第二语料库中各个词语进行匹配，得到所述分词结果包括的反映诉求情感的关键词组；根据词语与情感熵值的对应关系确定所述关键词组中各个关键词对应的情感熵值，所述情感熵值用于指示情感倾向程度；根据所述各个关键词对应的情感熵值确定所述目标对象的目标诉求情感信息。In an optional implementation manner, the
在一种可选的实施方式中，识别模块402根据所述各个关键词对应的情感熵值确定所述目标对象的目标诉求情感信息，具体为根据所述各个关键词对应的情感熵值计算得到情感熵值均值；将所述情感熵值均值确定为所述目标对象的目标诉求情感信息。In an optional implementation manner, the
在一种可选的实施方式中，处理模块403根据所述目标诉求情感信息和所述目标诉求意图确定目标诉求场景，具体为根据诉求意图与诉求场景的对应关系确定所述目标诉求意图对应的至少一个诉求场景；从所述至少一个诉求场景中筛选出与所述目标情感信息匹配的诉求场景；将所述匹配的诉求场景确定为目标诉求场景。In an optional implementation manner, the
在一种可选的实施方式中，处理模块403，还用于获取所述目标对象的对象数据；根据所述对象数据提取对象标签；处理模块403根据所述目标诉求场景从第一语料库中匹配出目标诉求处理方案，具体为从所述第一语料库中查询出所述目标诉求场景对应的至少一种诉求处理方案；从所述至少一种诉求处理方案中筛选出与所述对象标签匹配的诉求处理方案；将所述匹配的诉求处理方案确定为目标诉求处理方案。In an optional implementation manner, the
在一种可选的实施方式中，处理模块403，还用于从所述分词结果提取产品标签；根据所述产品标签确定产品类型；处理模块403根据所述目标诉求场景从第一语料库中匹配出目标诉求处理方案，具体为从所述第一语料库中查询出所述目标诉求场景对应的至少一种诉求处理方案；从所述至少一种诉求处理方案中筛选出与所述产品类型匹配的诉求处理方案；将所述匹配的诉求处理方案确定为目标诉求处理方案。In an optional implementation manner, the
在一种可选的实施方式中，处理模块403，还用于在根据所述目标诉求场景从第一语料库中匹配出目标诉求处理方案之后，将所述目标诉求处理方案发送至目标智能终端；当接收到目标智能终端根据所述目标诉求处理方案返回的确认指令时，执行所述按照所述目标诉求处理方案对所述目标对象进行诉求处理的步骤。In an optional implementation manner, the
可见，图4所示的实施例中，信息处理装置可以对目标对象的诉求文本进行分词处理，得到分词结果，然后对分词结果进行情感识别，得到目标对象的目标诉求情感信息，并对分词结果进行意图识别，得到目标对象的目标诉求意图，之后可以根据目标诉求情感信息和目标诉求意图确定目标诉求场景，并根据目标诉求场景从第一语料库中匹配出目标诉求处理方案，进而按照目标诉求处理方案进行诉求处理，该过程可以提高诉求处理效率和诉求处理质量。It can be seen that, in the embodiment shown in FIG. 4 , the information processing device can perform word segmentation processing on the appeal text of the target object to obtain a word segmentation result, and then perform emotion recognition on the word segmentation result to obtain the target appeal emotional information of the target object, and analyze the word segmentation result. Carry out intention recognition to obtain the target appeal intention of the target object, and then determine the target appeal scene according to the target appeal emotional information and target appeal intention, and match the target appeal processing plan from the first corpus according to the target appeal scene, and then process according to the target appeal This process can improve the efficiency and quality of appeal processing.
请参阅图5，为本申请实施例提供的一种计算机设备的结构示意图。本实施例中所描述的计算机设备可以包括：一个或多个处理器1000和存储器2000。处理器1000和存储器2000可以通过总线等方式连接。Please refer to FIG. 5 , which is a schematic structural diagram of a computer device according to an embodiment of the present application. The computer device described in this embodiment may include: one or
处理器1000可以是中央处理模块(Central Processing Unit，CPU)，该处理器还可以是其他通用处理器、数字信号处理器(Digital Signal Processor，DSP)、专用集成电路(Application Specific Integrated Circuit，ASIC)、现成可编程门阵列(Field-Programmable Gate Array，FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件等。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。The
存储器2000可以是高速RAM存储器，也可为非不稳定的存储器(non-volatilememory)，例如磁盘存储器。存储器2000用于存储一组程序代码，处理器1000可以调用存储器2000中存储的程序代码。具体地：The
处理器1000，用于获取目标对象的诉求文本，并对所述诉求文本进行分词处理，得到分词结果；对所述分词结果进行情感识别，得到所述目标对象的目标诉求情感信息，并对所述分词结果进行意图识别，得到所述目标对象的目标诉求意图；根据所述目标诉求情感信息和所述目标诉求意图确定目标诉求场景，并根据所述目标诉求场景从第一语料库中匹配出目标诉求处理方案，所述第一语料库包括多种诉求处理方案，每种诉求处理方案对应一种或多种诉求场景；按照所述目标诉求处理方案进行诉求处理。The
在一个实施例中，处理器1000，具体用于：In one embodiment, the
将所述分词结果中各个词语与第二语料库中各个词语进行匹配，得到所述分词结果包括的反映诉求情感的关键词组；Matching each word in the word segmentation result with each word in the second corpus to obtain a keyword group that reflects the emotion of the appeal included in the word segmentation result;
根据词语与情感熵值的对应关系确定所述关键词组中各个关键词对应的情感熵值，所述情感熵值用于指示情感倾向程度；Determine the sentiment entropy value corresponding to each keyword in the keyword group according to the correspondence between the word and the sentiment entropy value, and the sentiment entropy value is used to indicate the degree of sentiment tendency;
根据所述各个关键词对应的情感熵值确定所述目标对象的目标诉求情感信息。The target appeal emotional information of the target object is determined according to the emotional entropy value corresponding to each keyword.
在一个实施例中，在根据所述各个关键词对应的情感熵值确定所述目标对象的目标诉求情感信息时，处理器1000，具体用于：In one embodiment, when determining the target appeal emotional information of the target object according to the emotional entropy value corresponding to each keyword, the
根据所述各个关键词对应的情感熵值计算得到情感熵值均值；Calculate the mean value of emotional entropy according to the emotional entropy value corresponding to each keyword;
将所述情感熵值均值确定为所述目标对象的目标诉求情感信息。The mean value of the emotional entropy value is determined as the target appeal emotional information of the target object.
在一个实施例中，处理器1000，还具体用于：In one embodiment, the
根据诉求意图与诉求场景的对应关系确定所述目标诉求意图对应的至少一个诉求场景；Determine at least one appeal scene corresponding to the target appeal intent according to the correspondence between the appeal intent and the appeal scene;
从所述至少一个诉求场景中筛选出与所述目标情感信息匹配的诉求场景；Filter out the appeal scene matching the target emotional information from the at least one appeal scene;
将所述匹配的诉求场景确定为目标诉求场景。The matched appeal scenario is determined as the target appeal scenario.
在一个实施例中，处理器1000，还用于：In one embodiment, the
获取所述目标对象的对象数据；obtaining object data of the target object;
根据所述对象数据提取对象标签；extracting object tags according to the object data;
在根据所述目标诉求场景从第一语料库中匹配出目标诉求处理方案时，处理器1000，具体用于：When matching the target appeal processing plan from the first corpus according to the target appeal scenario, the
从所述第一语料库中查询出所述目标诉求场景对应的至少一种诉求处理方案；at least one appeal processing scheme corresponding to the target appeal scene is queried from the first corpus;
从所述至少一种诉求处理方案中筛选出与所述对象标签匹配的诉求处理方案；Screen out a claim processing solution matching the object label from the at least one claim processing solution;
将所述匹配的诉求处理方案确定为目标诉求处理方案。The matched appeal processing scheme is determined as the target appeal processing scheme.
在一个实施例中，处理器1000，还用于：In one embodiment, the
从所述分词结果提取产品标签；extracting product tags from the word segmentation results;
根据所述产品标签确定产品类型；Determine the product type from the product label;
在根据所述目标诉求场景从第一语料库中匹配出目标诉求处理方案时，处理器1000，还具体用于：When matching the target appeal processing plan from the first corpus according to the target appeal scenario, the
从所述第一语料库中查询出所述目标诉求场景对应的至少一种诉求处理方案；at least one appeal processing scheme corresponding to the target appeal scene is queried from the first corpus;
从所述至少一种诉求处理方案中筛选出与所述产品类型匹配的诉求处理方案；Screen out a claim processing solution matching the product type from the at least one claim processing solution;
将所述匹配的诉求处理方案确定为目标诉求处理方案。The matched appeal processing scheme is determined as the target appeal processing scheme.
在一个实施例中，在根据所述目标诉求场景从第一语料库中匹配出目标诉求处理方案之后，处理器1000，还用于：In one embodiment, after matching the target appeal processing scheme from the first corpus according to the target appeal scenario, the
将所述目标诉求处理方案发送至目标智能终端；sending the target appeal processing plan to the target intelligent terminal;
当接收到目标智能终端根据所述目标诉求处理方案返回的确认指令时，执行所述按照所述目标诉求处理方案对所述目标对象进行诉求处理的步骤。When receiving the confirmation instruction returned by the target intelligent terminal according to the target appeal processing scheme, the step of performing appeal processing on the target object according to the target appeal processing scheme is performed.
具体实现中，本申请实施例中所描述的处理器1000可执行图2实施例、图3实施例所描述的实现方式，也可执行本申请实施例所描述的实现方式，在此不再赘述。In the specific implementation, the
在本申请各个实施例中的各功能模块可以集成在一个处理模块中，也可以是各个模块单独物理存在，也可以是两个或两个以上模块集成在一个模块中。上述集成的模块既可以采样硬件的形式实现，也可以采样软件功能模块的形式实现。Each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically alone, or two or more modules may be integrated into one module. The above-mentioned integrated modules can be implemented in the form of sampling hardware or in the form of sampling software function modules.
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程，是可以通过计算机程序来指令相关的硬件来完成，所述的程序可存储于一计算机可读取存储介质中，该程序在执行时，可包括如上述各方法的实施例的流程。其中，所述的计算机可读存储介质可为易失性的或非易失性的。例如，该计算机存储介质可以为磁碟、光盘、只读存储记忆体(Read-Only Memory，ROM)或随机存储记忆体(Random Access Memory，RAM)等。所述的计算机可读存储介质可主要包括存储程序区和存储数据区，其中，存储程序区可存储操作系统、至少一个功能所需的应用程序等；存储数据区可存储根据区块链节点的使用所创建的数据等。Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be implemented by instructing relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium. During execution, the processes of the embodiments of the above-mentioned methods may be included. Wherein, the computer-readable storage medium can be volatile or non-volatile. For example, the computer storage medium may be a magnetic disk, an optical disk, a read-only memory (Read-Only Memory, ROM), or a random access memory (Random Access Memory, RAM) or the like. The computer-readable storage medium may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function, and the like; Use the created data, etc.
其中，本申请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。Among them, the blockchain referred to in this application is a new application mode of computer technology such as distributed data storage, point-to-point transmission, consensus mechanism, and encryption algorithm. Blockchain, essentially a decentralized database, is a series of data blocks associated with cryptographic methods. Each data block contains a batch of network transaction information 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.
以上所揭露的仅为本申请一种较佳实施例而已，当然不能以此来限定本申请之权利范围，本领域普通技术人员可以理解实现上述实施例的全部或部分流程，并依本申请权利要求所作的等同变化，仍属于本申请所涵盖的范围。What is disclosed above is only a preferred embodiment of the present application, and of course, it cannot limit the scope of the right of the present application. Those skilled in the art can understand that all or part of the process of implementing the above-mentioned embodiment can be realized according to the right of the present application. The equivalent changes required to be made still fall within the scope covered by this application.
Claims (10)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210314713.XACN114579750A (en) | 2022-03-22 | 2022-03-22 | Information processing method and device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210314713.XACN114579750A (en) | 2022-03-22 | 2022-03-22 | Information processing method and device, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114579750Atrue CN114579750A (en) | 2022-06-03 |
Family
ID=81776429
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210314713.XAPendingCN114579750A (en) | 2022-03-22 | 2022-03-22 | Information processing method and device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114579750A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115081843A (en)* | 2022-06-13 | 2022-09-20 | 深圳供电局有限公司 | Intelligent early warning method and system based on complaint risk |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105893344A (en)* | 2016-03-28 | 2016-08-24 | 北京京东尚科信息技术有限公司 | User semantic sentiment analysis-based response method and device |
| CN113342960A (en)* | 2021-07-07 | 2021-09-03 | 上海华客信息科技有限公司 | Client appeal processing method, system, device and storage medium |
- 2022
- 2022-03-22CNCN202210314713.XApatent/CN114579750A/enactivePending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105893344A (en)* | 2016-03-28 | 2016-08-24 | 北京京东尚科信息技术有限公司 | User semantic sentiment analysis-based response method and device |
| CN113342960A (en)* | 2021-07-07 | 2021-09-03 | 上海华客信息科技有限公司 | Client appeal processing method, system, device and storage medium |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115081843A (en)* | 2022-06-13 | 2022-09-20 | 深圳供电局有限公司 | Intelligent early warning method and system based on complaint risk |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10997258B2 (en) | Bot networks | |
| CN110929125B (en) | Search recall method, device, equipment and storage medium thereof | |
| CN110168535B (en) | An information processing method, terminal, and computer storage medium | |
| EP3652685A1 (en) | System and method for identifying and providing personalized self-help content with artificial intelligence in a customer self-help system | |
| CN112699645B (en) | Corpus labeling method, apparatus and device | |
| CN104346480A (en) | Information mining method and device | |
| CN114528851B (en) | Reply sentence determination method, reply sentence determination device, electronic equipment and storage medium | |
| CN107436916B (en) | Intelligent answer prompting method and device | |
| CN107016132A (en) | Online question bank quality improving method and system and terminal equipment | |
| CN114202203A (en) | Complaint work order processing method, device, storage medium and electronic equipment | |
| CN113553431B (en) | User tag extraction method, device, equipment and medium | |
| CN112288584A (en) | Insurance application processing method and device, computer readable medium and electronic equipment | |
| CN113010664B (en) | Data processing method, device and computer equipment | |
| WO2023272862A1 (en) | Risk control recognition method and apparatus based on network behavior data, and electronic device and medium | |
| CN114328913A (en) | Text classification method and device, computer equipment and storage medium | |
| CN110705308A (en) | Method and device for recognizing field of voice information, storage medium and electronic equipment | |
| CN118094239A (en) | Image and text rating method, device and computer-readable storage medium | |
| CN113435859A (en) | Letter processing method and device, electronic equipment and computer readable medium | |
| CN116562837A (en) | Person-post matching method, device, electronic device, and computer-readable storage medium | |
| US20140095527A1 (en) | Expanding high level queries | |
| CN119739987A (en) | Label system construction, label providing method, device, equipment and storage medium | |
| CN114168837A (en) | Chatbot searching method, equipment and storage medium | |
| CN114579750A (en) | Information processing method and device, computer equipment and storage medium | |
| CN117009528A (en) | Business processing method, device, equipment and medium based on natural language processing | |
| CN111782776B (en) | A method and device for realizing intention recognition by slot filling |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | Application publication date:20220603 | |
| WD01 | Invention patent application deemed withdrawn after publication |