技术领域technical field
本公开涉及大数据领域，特别涉及一种数据处理方法、装置以及计算机可读存储介质。The present disclosure relates to the field of big data, and in particular to a data processing method, device and computer-readable storage medium.
背景技术Background technique
电子商务行业经过多年的发展已经非常成熟，多数电子商务网站都具备完善的商品，库存，订单和售后体系。但是，电子商务网站的海量商品却给用户造成了巨大的困扰，使用户难以抉择。商品问答是一种较新的形式，答案可以从消费者关心的角度对商品进行概括，描述商品的全貌，并对消费者提供一些购买上的指导意见。After years of development, the e-commerce industry has become very mature, and most e-commerce websites have a complete product, inventory, order and after-sales system. However, the massive amount of commodities on the e-commerce website has caused huge troubles to users, making it difficult for users to make a choice. Commodity question and answer is a relatively new form. The answer can summarize the commodity from the perspective of consumers' concern, describe the whole picture of the commodity, and provide consumers with some purchasing guidance.
目前，商品问答过程中包括提问者和回答者。提问者：商品购买意向者，关注某个商品或品类，但是仍处于犹豫中的用户，提出自己关心的问题。回答者：例如购买过某件商品的用户，对提问者提出的问题给出个人解答。提问者可以根据回答者的解答选择是否购买商品。Currently, the product Q&A process includes both a questioner and an answerer. Questioner: Those who are interested in buying a product, users who pay attention to a certain product or category but are still hesitating, ask questions that they care about. Respondent: For example, a user who has purchased a product and gives a personal answer to the question raised by the questioner. The questioner can choose whether to purchase the product according to the answer of the respondent.
发明内容Contents of the invention
发明人发现：上述商品问答的形式中问题的回复率，问答的用户体验严重依赖回答者给出的答案。而大多数商品问答的答案难以完全覆盖提问者提出的全部问题，回复率不高，并且回复时间较长，导致提问者不能及时获得答复，降低了提问者的体验。The inventor found that: the reply rate of questions in the form of product question-and-answer and the user experience of question-and-answer heavily depend on the answers given by the respondent. However, the answers to most product questions and answers are difficult to completely cover all the questions raised by the questioners, the response rate is not high, and the response time is long, which prevents the questioners from getting answers in time and reduces the experience of the questioners.
本公开所要解决的一个技术问题是：如何提高商品问答中答复的效率，提升用户体验。A technical problem to be solved in the present disclosure is: how to improve the efficiency of answering product questions and answers and improve user experience.
根据本公开的一些实施例，提供的一种数据处理方法，包括：获取用户关于对象的提问信息；查找已存储的对象的评论信息；根据评论信息和提问信息的关键词、语法信息和语义信息中的至少一项信息，将评论信息与提问信息进行匹配；根据匹配结果向用户推荐至少一条评论信息，作为用户提问的答案。According to some embodiments of the present disclosure, a data processing method is provided, including: obtaining user's question information about an object; searching for stored object comment information; keywords, grammatical information, and semantic information based on the comment information and question information Match at least one piece of information in the comment information with the question information; recommend at least one piece of comment information to the user according to the matching result as the answer to the user's question.
在一些实施例中，将评论信息与提问信息进行匹配包括：分别确定各条评论信息与提问信息的关键词匹配度、语法匹配度和语义匹配度；将同一条评论信息对应的关键词匹配度、语法匹配度和语义匹配度进行加权，作为该条评论信息与提问信息的匹配度。In some embodiments, matching the comment information with the question information includes: respectively determining the keyword matching degree, grammatical matching degree and semantic matching degree of each piece of comment information and the question information; matching the keyword matching degree corresponding to the same piece of comment information , grammatical matching degree and semantic matching degree are weighted as the matching degree of the comment information and the question information.
在一些实施例中，根据匹配结果向用户推荐至少一条评论信息包括：根据各条评论信息对应的用户信用等级、用户注册信息、评论时间信息和采用率中的至少一项信息，对匹配结果进行修正；根据修正后的匹配结果向用户推荐至少一条评论信息。In some embodiments, recommending at least one piece of comment information to the user according to the matching result includes: performing an operation on the matching result according to at least one item of information corresponding to each piece of comment information, including user credit rating, user registration information, comment time information, and adoption rate. Amendment; recommending at least one piece of comment information to the user according to the amended matching result.
在一些实施例中，匹配结果包括各条评论信息与提问信息的匹配度；对匹配结果进行修正包括：将评论信息对应的用户信用等级权重、用户注册权重、评论时间权重和答案采用率权重中的至少一项与该评论信息对应的匹配度相乘，得到的乘积作为修正后的匹配结果；其中，用户信用等级越高，用户信用等级权重越高；用户注册时间越早，注册权重越高；评论时间与提问信息对应的时间差距越小，评论时间权重越高。In some embodiments, the matching result includes the matching degree of each piece of comment information and the question information; modifying the matching result includes: adding the user credit rating weight, user registration weight, comment time weight and answer adoption rate weight corresponding to the comment information Multiply at least one of the matching degree corresponding to the review information, and the obtained product is used as the corrected matching result; among them, the higher the user credit rating, the higher the user credit rating weight; the earlier the user registration time, the higher the registration weight ; The smaller the time difference between the comment time and the question information, the higher the weight of the comment time.
在一些实施例中，该方法还包括：获取评论信息，评论信息包括从对象评论页面、问答页面、社区类页面和客服系统中至少一处获取的评论信息；建立评论信息与对象信息、用户信息和时间信息的对应关系并存储。In some embodiments, the method further includes: obtaining comment information, the comment information including comment information obtained from at least one of object comment pages, question and answer pages, community pages and customer service systems; establishing comment information and object information, user information Correspondence with time information and stored.
在一些实施例中，根据关键词将提问信息与评论信息进行匹配包括：对提问信息和评论信息分别进行分词；根据各个词语的词频和包含该词语的训练样本数，分别提取提问信息和评论信息的关键词；根据各条评论信息与提问信息的关键词的相似度，确定各条评论信息与提问信息的关键词匹配度。In some embodiments, matching the question information with the comment information according to the keywords includes: segmenting the question information and the comment information respectively; extracting the question information and the comment information respectively according to the word frequency of each word and the number of training samples containing the word keyword; determine the keyword matching degree of each piece of comment information and the question information according to the similarity of each piece of comment information and the keyword of the question information.
在一些实施例中，根据语法信息将提问信息与评论信息进行匹配包括：对提问信息和评论信息分别进行分词；根据提问信息和评论信息中各个词语的词性，分别确定提问信息和评论信息中各个句子的语法结构；根据提问信息和评论信息中各个句子的语法结构的相似度，确定各条评论信息与提问信息的语法匹配度。In some embodiments, matching the question information with the comment information according to the grammatical information includes: segmenting the question information and the comment information respectively; The grammatical structure of the sentence; according to the similarity of the grammatical structure of each sentence in the question information and the comment information, determine the grammatical matching degree between each piece of comment information and the question information.
在一些实施例中，根据语义信息将提问信息与评论信息进行匹配包括：对提问信息和评论信息分别进行分词；分别确定提问信息和评论信息中各个词语的词向量；根据提问信息和评论信息中各个词语的词向量，计算各条评论信息与提问信息的语义匹配度。In some embodiments, matching the question information with the comment information according to the semantic information includes: segmenting the question information and the comment information respectively; respectively determining the word vectors of each word in the question information and the comment information; The word vector of each word calculates the semantic matching degree of each comment information and question information.
根据本公开的另一些实施例，提供的一种数据处理装置，包括：提问信息获取模块，用于获取用户关于对象的提问信息；评论信息查找模块，用于查找已存储的对象的评论信息；匹配模块，用于根据评论信息和提问信息的关键词、语法信息和语义信息中的至少一项信息，将评论信息与提问信息进行匹配；答案推荐模块，用于根据匹配结果向用户推荐至少一条评论信息，作为用户提问的答案。According to some other embodiments of the present disclosure, a data processing device is provided, including: a question information acquisition module, configured to acquire user's question information about an object; a comment information search module, configured to search for stored comment information of an object; The matching module is used to match the comment information with the question information according to at least one of the keywords, grammatical information and semantic information of the comment information and the question information; the answer recommendation module is used to recommend at least one item to the user according to the matching result. Comment information, as an answer to a user's question.
在一些实施例中，匹配模块用于分别确定各条评论信息与提问信息的关键词匹配度、语法匹配度和语义匹配度，将同一条评论信息对应的关键词匹配度、语法匹配度和语义匹配度进行加权，作为该条评论信息与提问信息的匹配度。In some embodiments, the matching module is used to respectively determine the keyword matching degree, grammatical matching degree and semantic matching degree of each piece of comment information and question information, and the corresponding keyword matching degree, grammatical matching degree and semantic matching degree of the same comment information The matching degree is weighted as the matching degree between the comment information and the question information.
在一些实施例中，答案推荐模块用于根据各条评论信息对应的用户信用等级、用户注册信息、评论时间信息和采用率中的至少一项信息，对匹配结果进行修正，根据修正后的匹配结果向用户推荐至少一条评论信息。In some embodiments, the answer recommendation module is used to modify the matching result according to at least one item of user credit rating, user registration information, comment time information and adoption rate corresponding to each piece of comment information, and according to the corrected matching As a result, at least one piece of review information is recommended to the user.
在一些实施例中，匹配结果包括各条评论信息与提问信息的匹配度；答案推荐模块用于将评论信息对应的用户信用等级权重、用户注册权重、评论时间权重和答案采用率权重中的至少一项与该评论信息对应的匹配度相乘，得到的乘积作为修正后的匹配结果；其中，用户信用等级越高，用户信用等级权重越高；用户注册时间越早，注册权重越高；评论时间与提问信息对应的时间差距越小，评论时间权重越高。In some embodiments, the matching result includes the matching degree of each piece of comment information and the question information; the answer recommendation module is used to use at least One item is multiplied by the matching degree corresponding to the comment information, and the obtained product is used as the corrected matching result; among them, the higher the user credit rating, the higher the weight of the user credit rating; the earlier the user registration time, the higher the registration weight; the comment The smaller the time difference between the time and the question information, the higher the weight of the comment time.
在一些实施例中，该装置还包括：评论信息处理模块，用于获取评论信息，建立评论信息与对象信息、用户信息和时间信息的对应关系并存储；其中，评论信息包括从对象评论页面、问答页面、社区类页面和客服系统中至少一处获取的评论信息。In some embodiments, the device further includes: a comment information processing module, configured to acquire comment information, establish and store the corresponding relationship between comment information and object information, user information, and time information; wherein, the comment information includes comments from the object comment page, Comment information obtained from at least one of the Q&A page, community page, and customer service system.
在一些实施例中，匹配模块用于对提问信息和评论信息分别进行分词，根据各个词语的词频和包含该词语的训练样本数，分别提取提问信息和评论信息的关键词，根据各条评论信息与提问信息的关键词的相似度，确定各条评论信息与提问信息的关键词匹配度。In some embodiments, the matching module is used to segment the question information and the comment information respectively, extract the keywords of the question information and the comment information respectively according to the word frequency of each word and the number of training samples containing the word, and extract the keywords of the question information and comment information according to each piece of comment information The similarity with the keyword of the question information determines the matching degree of the keyword between each piece of comment information and the question information.
在一些实施例中，匹配模块用于对提问信息和评论信息分别进行分词，根据提问信息和评论信息中各个词语的词性，分别确定提问信息和评论信息中各个句子的语法结构，根据提问信息和评论信息中各个句子的语法结构的相似度，确定各条评论信息与提问信息的语法匹配度。In some embodiments, the matching module is used to segment the question information and comment information respectively, determine the grammatical structure of each sentence in the question information and comment information according to the part of speech of each word in the question information and comment information, and determine the grammatical structure of each sentence in the question information and comment information according to the question information and The similarity of the grammatical structure of each sentence in the comment information determines the grammatical matching degree of each comment information and the question information.
在一些实施例中，匹配模块用于对提问信息和评论信息分别进行分词，分别确定提问信息和评论信息中各个词语的词向量，根据提问信息和评论信息中各个词语的词向量，计算各条评论信息与提问信息的语义匹配度。In some embodiments, the matching module is used to separately segment the question information and comment information, respectively determine the word vectors of each word in the question information and comment information, and calculate the word vectors of each word in the question information and comment information. Semantic matching degree of comment information and question information.
根据本公开的又一些实施例，提供的一种数据处理装置，包括：存储器；以及耦接至存储器的处理器，处理器被配置为基于存储在存储器设备中的指令，执行如前述任意实施例的数据处理方法。According to some other embodiments of the present disclosure, there is provided a data processing apparatus, including: a memory; and a processor coupled to the memory, the processor is configured to execute any of the foregoing embodiments based on instructions stored in the memory device. data processing method.
根据本公开的再一些实施例，提供的一种计算机可读存储介质，其上存储有计算机程序，其中，该程序被处理器执行时实现前述任意实施例的数据处理方法的步骤。According to some further embodiments of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored, wherein, when the program is executed by a processor, the steps of the data processing method in any of the foregoing embodiments are implemented.
本公开针对用户关于对象的提问信息，将该对象的评论信息与提问信息进行匹配，根据匹配结果选取至少一条评论信息作为提问的答案推荐给用户。由于评论信息中包含其他用户关于该对象的评价，利用评论信息可以及时高效地为提问者推荐与问题相关的答案，提高了商品问答中答复的效率，提升用户体验。The present disclosure matches the user's question information about an object, matches the object's comment information with the question information, and selects at least one comment information as an answer to the question according to the matching result and recommends it to the user. Since the comment information contains other users' evaluations about the object, the comment information can be used to recommend answers related to the question for the questioner in a timely and efficient manner, which improves the efficiency of answering product questions and answers and improves user experience.
通过以下参照附图对本公开的示例性实施例的详细描述，本公开的其它特征及其优点将会变得清楚。Other features of the present disclosure and advantages thereof will become apparent through the following detailed description of exemplary embodiments of the present disclosure with reference to the accompanying drawings.
附图说明Description of drawings
为了更清楚地说明本公开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公开的一些实施例，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 technical solutions in the embodiments of the present disclosure or the prior art, the following will briefly introduce the drawings that need to be used in 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 disclosure. For those skilled in the art, other drawings can also be obtained according to these drawings without creative work.
图1示出本公开的一些实施例的数据处理方法的流程示意图。Fig. 1 shows a schematic flowchart of a data processing method in some embodiments of the present disclosure.
图2示出本公开的另一些实施例的数据处理方法的流程示意图。Fig. 2 shows a schematic flowchart of a data processing method in some other embodiments of the present disclosure.
图3示出本公开的一些实施例的数据处理装置的结构示意图。Fig. 3 shows a schematic structural diagram of a data processing device in some embodiments of the present disclosure.
图4示出本公开的另一些实施例的数据处理装置的结构示意图。Fig. 4 shows a schematic structural diagram of a data processing device according to other embodiments of the present disclosure.
图5示出本公开的又一些实施例的数据处理装置的结构示意图。Fig. 5 shows a schematic structural diagram of a data processing device in some other embodiments of the present disclosure.
具体实施方式Detailed ways
下面将结合本公开实施例中的附图，对本公开实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本公开一部分实施例，而不是全部的实施例。以下对至少一个示例性实施例的描述实际上仅仅是说明性的，决不作为对本公开及其应用或使用的任何限制。基于本公开中的实施例，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，都属于本公开保护的范围。The following will clearly and completely describe the technical solutions in the embodiments of the present disclosure with reference to the accompanying drawings in the embodiments of the present disclosure. Apparently, the described embodiments are only some of the embodiments of the present disclosure, not all of them. The following description of at least one exemplary embodiment is merely illustrative in nature and in no way intended as any limitation of the disclosure, its application or uses. Based on the embodiments in the present disclosure, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present disclosure.
本公开提供一种数据处理方法，可以用于商品问答场景。下面结合图1描述本公开数据处理方法的一些实施例。The present disclosure provides a data processing method that can be used in a commodity question answering scenario. Some embodiments of the data processing method of the present disclosure are described below with reference to FIG. 1 .
图1为本公开数据处理方法一些实施例的流程图。如图1所示，该实施例的方法包括：步骤S102～S108。Fig. 1 is a flowchart of some embodiments of the data processing method of the present disclosure. As shown in FIG. 1 , the method of this embodiment includes: steps S102-S108.
在步骤S102，获取用户关于对象的提问信息。In step S102, the user's question information about the object is acquired.
对于商品问答的场景，用户可以在商品详情页面的提问区进行提问，系统则可以自动获取提问信息所针对的对象信息，例如商品编号等。用户还可以在电子商务平台的客服系统或社区论坛等位置发出提问，在这种用户没有明确指出提问对象的情况下，可以对提问语句进行处理提取对象信息。例如，可以对提问语句进行分词，清洗(例如去除停用词)，将各个词语与词库中的词语进行比对，确定提问信息对应的对象。也可以采取其他方法确定提问信息对应的对象，不限于所举示例。对象可以是具体的商品，也可以是品类，例如手机，也可以是某一品牌的产品，例如小米手机，可以根据实际需求确定。For product question-and-answer scenarios, users can ask questions in the question area on the product details page, and the system can automatically obtain information about the object of the question, such as the product number. Users can also ask questions in the customer service system of the e-commerce platform or community forums. In the case where the user does not clearly indicate the object of the question, the query statement can be processed to extract the object information. For example, it is possible to segment the question sentence, clean it (for example, remove stop words), compare each word with the words in the thesaurus, and determine the object corresponding to the question information. Other methods may also be used to determine the object corresponding to the question information, which is not limited to the examples given. The object can be a specific commodity, or a category, such as a mobile phone, or a product of a certain brand, such as a Xiaomi mobile phone, which can be determined according to actual needs.
在步骤S104，查找已存储的对象的评论信息。In step S104, the stored comment information of the object is searched.
评论信息可以是用户产生的关于对象的任何形式的描述信息。例如，从对象评论页面、问答页面、社区类页面和客服系统中至少一处获取的评论信息。例如，用户使用商品后可以在评论页面发表评论，可以回答其他用户的提问，或者在论坛等社区类页面发表测评文章，还可以向客服反映使用中遇到的问题或者要求售后服务等。这些评论信息可以反映用户使用对象的体验，实际上发表过评论信息的用户在回答其他用户关于对象的使用体验的问题时，大多数回答与评论信息相似。因此评论信息可以作为提问答案反馈给用户。Comment information can be any form of descriptive information about objects generated by users. For example, comment information obtained from at least one of object comment pages, question and answer pages, community pages, and customer service systems. For example, after using the product, users can post comments on the review page, answer questions from other users, or publish evaluation articles on community pages such as forums, and report problems encountered in use to customer service or request after-sales service. These comment information can reflect the user's experience of using the object. In fact, when the users who have posted the comment information answer other users' questions about the use experience of the object, most of the answers are similar to the comment information. Therefore, the comment information can be fed back to the user as the answer to the question.
评论信息可以存储于数据库，每隔预设周期进行更新。评论信息一般针对特定的对象，可以直接确定评论信息对应的对象。进一步，可以根据对象获取对象品类、品牌等信息。可以将对象信息(包括品类、品牌等)与评论信息进行关联存储。对于获取的评论信息可以进行预处理，例如，对评论信息进行自动审核，去除其中包含敏感词、或者与对象无关、不属于正常评论信息的评论信息等。自动审核的过程可以采用自然语言处理中语义识别的方法，属于现有技术，在此不再赘述。Comment information can be stored in the database and updated every preset period. The comment information is generally aimed at a specific object, and the object corresponding to the comment information can be directly determined. Further, information such as object category and brand may be obtained according to the object. Object information (including category, brand, etc.) and comment information can be associated and stored. The obtained comment information can be preprocessed, for example, the comment information is automatically reviewed, and the comment information that contains sensitive words, or has nothing to do with the object, and does not belong to normal comment information, etc. is removed. The process of automatic review can adopt the method of semantic recognition in natural language processing, which belongs to the prior art and will not be repeated here.
进一步，可以将评论信息进行分词、清洗(例如去除停用词)、提取关键词、确定句子语法、确定语义信息(例如情感、态度等)等处理。可以用于后续与提问信息匹配，具体方法将在后续进行描述。至此，数据库中可以存储对象信息、评论信息的关键词、句子语法、语义信息以及分词结果等信息。Further, comment information can be segmented, cleaned (such as removing stop words), keywords are extracted, sentence grammar is determined, semantic information (such as emotion, attitude, etc.) is determined, etc. It can be used for subsequent matching with question information, and the specific method will be described later. So far, the database can store information such as object information, keywords of comment information, sentence grammar, semantic information, and word segmentation results.
将提问信息对应的对象信息与数据库中存储的对象信息进行比对，可以确定关于提问对象的评论信息。Comparing the object information corresponding to the question information with the object information stored in the database, the comment information about the question object can be determined.
在步骤S106，根据评论信息和提问信息的关键词、语法信息和语义信息中的至少一项信息，将评论信息与提问信息进行匹配。In step S106, the comment information is matched with the question information according to at least one item of information among keywords, grammatical information, and semantic information of the comment information and the question information.
在一些实施例中，根据关键词将评论信息与提问信息进行匹配可以采用以下方法：对提问信息和评论信息分别进行分词；根据各个词语的词频和包含该词语的训练样本数，分别提取提问信息和评论信息的关键词；根据各条评论信息与提问信息的关键词的相似度，确定各条评论信息与提问信息的关键词匹配度。In some embodiments, the following methods may be used to match the comment information and the question information according to the keywords: segment the question information and the comment information respectively; extract the question information according to the word frequency of each word and the number of training samples containing the word and the keywords of the comment information; according to the similarity of the keywords of each piece of comment information and the question information, determine the matching degree of the keywords of each piece of comment information and the question information.
如果评论信息在存储时已经进行了分词、提取关键词等处理，则在该步骤可以直接应用。可以利用TF-IDF(Term Frequency–Inverse Document Frequency，词频-逆文档频率)算法提取评论信息或提问信息的关键词。具体的，计算各个词语的词频与逆文档频率的乘积，得到词语的重要程度，根据重要程度选取关键词。词语的词频例如为该词在文本中出现的次数与文本总词数的比值。词语的逆文档频率例如为训练文本的总数与包含该词语的文本数的比值的对数值。关键词的提取还可以采取其他算法，例如RAKE(Rapid AutomaticKeyword Extraction，快速自动关键词提取)等算法，不限于所举示例。If word segmentation and keyword extraction have been performed on the comment information during storage, it can be directly applied in this step. The TF-IDF (Term Frequency–Inverse Document Frequency) algorithm can be used to extract keywords of comment information or question information. Specifically, the product of the word frequency of each word and the inverse document frequency is calculated to obtain the importance of the word, and keywords are selected according to the importance. The word frequency of a word is, for example, the ratio of the number of times the word appears in the text to the total number of words in the text. The inverse document frequency of a word is, for example, the logarithm of the ratio of the total number of training texts to the number of texts containing the word. Keyword extraction may also adopt other algorithms, such as RAKE (Rapid Automatic Keyword Extraction, fast automatic keyword extraction) and other algorithms, which are not limited to the examples given.
对于评论信息或提问信息，当分词清洗后剩余的词语数量低于阈值的情况下，可以将所有词语作为关键词，不再使用提取关键词的算法。将提问信息中的关键词与每条评论信息的中各个关键词进行相似度的比对确定提问信息和评论信息的关键词匹配度。具体的，可以采用word2vector算法将关键词转换为词向量，通过计算提问信息中的关键词与评论信息的关键词的词向量的距离，确定两者的相似度，将提问信息中各个关键词对应的相似度相加可以得到提问信息与评论信息的关键词匹配度。通过上述方法计算关键词的词向量可以增加匹配的准确性，例如，小米手机中的“小米”一词和食用的小米中“小米”一词虽然是同一词，但含义不同，通过词向量判断两者的相似度很低可以进行区分。For comment information or question information, when the number of remaining words after word segmentation and cleaning is lower than the threshold, all words can be used as keywords, and the algorithm for extracting keywords can no longer be used. Comparing the similarity between the keywords in the question information and the keywords in each piece of comment information to determine the keyword matching degree between the question information and the comment information. Specifically, the word2vector algorithm can be used to convert keywords into word vectors, and by calculating the distance between the keywords in the question information and the word vectors of the keywords in the comment information, the similarity between the two is determined, and the corresponding keywords in the question information Adding the similarity of the query information and the comment information can get the keyword matching degree. Calculating the word vector of the keyword by the above method can increase the accuracy of matching. For example, although the word "millet" in the Xiaomi mobile phone and the word "millet" in the edible millet are the same word, they have different meanings. Judging by the word vector The similarity between the two is low enough to distinguish them.
在一些实施例中，根据语法信息将评论信息与提问信息进行匹配可以采用以下方法：对提问信息和评论信息分别进行分词；根据提问信息和评论信息中各个词语的词性，分别确定提问信息和评论信息中各个句子的语法结构；根据提问信息和评论信息中各个句子的语法结构的相似度，确定各条评论信息与提问信息的语法匹配度。In some embodiments, the following methods may be used to match the comment information and the question information according to the grammatical information: segment the question information and the comment information respectively; The grammatical structure of each sentence in the information; according to the similarity of the grammatical structure of each sentence in the question information and the comment information, determine the grammatical matching degree between each piece of comment information and the question information.
如果评论信息在存储时已经进行了分词、确定各个句子的语法信息等处理，则在该步骤可以直接应用。词语的词性例如名词、动词、形容词等。根据各个词语的顺序和词性可以确定句子的语法结构。通过比对语法结构，可以确定评论信息与提问信息的语法匹配度。可以根据现有的算法确定词性、语法结构以及评论信息与提问信息的语法匹配度，例如，语法树，在此不再赘述。语法结构除了各个词语的词性和顺序还可以包括句子中的各个词语，即比对句子的匹配度时可以匹配词语、词性以及顺序等特征。If the comment information has been processed such as word segmentation and determination of grammatical information of each sentence during storage, it can be directly applied in this step. The part of speech of words such as noun, verb, adjective, etc. According to the order and part of speech of each word, the grammatical structure of the sentence can be determined. By comparing the grammatical structures, the grammatical matching degree between the comment information and the question information can be determined. The part of speech, grammatical structure, and grammatical matching degree between comment information and question information can be determined according to existing algorithms, for example, syntax trees, which will not be repeated here. In addition to the part of speech and order of each word, the grammatical structure can also include each word in the sentence, that is, when comparing the matching degree of the sentence, features such as words, parts of speech, and order can be matched.
在一些实施例中，根据语义信息将评论信息与提问信息进行匹配可以采用以下方法：对提问信息和评论信息分别进行分词；分别确定提问信息和评论信息中各个词语的词向量；根据提问信息和评论信息中各个词语的词向量，计算各条评论信息与提问信息的语义匹配度。In some embodiments, the following methods may be used to match the comment information and the question information according to the semantic information: segment the question information and the comment information respectively; respectively determine the word vectors of each word in the question information and the comment information; The word vector of each word in the comment information, and calculate the semantic matching degree between each comment information and the question information.
如果评论信息在存储时已经进行了分词，在本步骤可以直接应用。确定词语的词向量可以采用word2vector算法，也可以采用其他算法，在此不再赘述。可以利用深度学习神经网络计算提问信息和评论信息的词向量矩阵的相似度。例如，可以采用CNN(Convolutional Neural Network，卷积神经网络)或RNN(Recurrent Neural Networks，循环神经网络)等计算评论信息与提问信息的语义匹配度，在此不再赘述。If the comment information has been word-segmented during storage, it can be directly applied in this step. The word2vector algorithm or other algorithms may be used to determine the word vector of the word, which will not be repeated here. The similarity of the word vector matrix of the question information and the comment information can be calculated by using the deep learning neural network. For example, CNN (Convolutional Neural Network, Convolutional Neural Network) or RNN (Recurrent Neural Networks, Recurrent Neural Network) can be used to calculate the semantic matching degree between comment information and question information, which will not be repeated here.
上述实施例的一种应用场景例如，用户想要购买苹果的一款手机，但是想了解其性能，可以发出提问，例如，手机使用是否流畅，有什么问题，照相是否清晰等等。针对用户发出的提问，系统自动搜索相匹配的购买用户在商品下的评论、论坛中的测评、社区中的用户讨论信息以及购买用户与客服咨询的信息等等，选取与用户提问最相关的若干评论返回给提问用户，则可以使提问用户在没有人回答提问的情况下及时了解手机情况，提升用户体验。An application scenario of the above embodiment. For example, the user wants to buy a mobile phone from Apple, but wants to know its performance, and can ask questions, such as whether the mobile phone is smooth to use, what is the problem, whether the camera is clear, etc. In response to the questions sent by users, the system automatically searches for matching purchase users' comments under the product, evaluations in forums, user discussion information in the community, and information about purchase users and customer service consultation, etc., and selects the most relevant ones for the user's questions. Comments are returned to the user who asked the question, so that the user can keep abreast of the situation of the mobile phone when no one answers the question, and improve the user experience.
上述实施例中的关键词匹配方法、语法匹配方法和语义匹配方法可以单独使用，也可以任意两种或三种结合使用。例如，分别确定各条评论信息与提问信息的关键词匹配度、语法匹配度和语义匹配度；将同一条评论信息对应的关键词匹配度、语法匹配度和语义匹配度进行加权，作为该条评论信息与提问信息的匹配度。可以根据实际测试匹配的准确率，分别为关键词匹配、语法匹配和语义匹配的结果设置不同的权重，将三种匹配度进行加权，作为评论信息与提问信息的匹配度。The keyword matching method, the syntax matching method and the semantic matching method in the above embodiments can be used alone, or any two or three can be used in combination. For example, determine the keyword matching degree, grammatical matching degree and semantic matching degree of each comment information and question information respectively; weight the keyword matching degree, grammatical matching degree and semantic matching degree corresponding to the same comment information as the The matching degree of comment information and question information. According to the accuracy of actual test matching, different weights can be set for the results of keyword matching, grammatical matching and semantic matching, and the three matching degrees can be weighted as the matching degree of comment information and question information.
在步骤S108，根据匹配结果向用户推荐至少一条评论信息，作为用户提问的答案。In step S108, at least one piece of comment information is recommended to the user according to the matching result as an answer to the user's question.
可以选取匹配度高于阈值的评论信息作为答案推荐给提问的用户。进一步，针对匹配度高于阈值的评论信息还可以进行处理，例如，可以对这些评论信息进行匹配，将内容相似度高于阈值的评论信息选取部分推荐给用户或者将评论信息根据含义进行分组推荐给用户。The comment information whose matching degree is higher than the threshold can be selected as the answer and recommended to the user asking the question. Further, the review information with a matching degree higher than the threshold can also be processed. For example, these review information can be matched, and the selected part of the review information whose content similarity is higher than the threshold is recommended to the user or the review information is grouped and recommended according to the meaning. to the user.
具体的，可以通过语法匹配选取含义相反或相对的评论信息分组推荐给用户，语法匹配参考前述实施例。例如，用户问题为小米手机和苹果手机哪个更好用，一些评论信息为小米手机比苹果手机好用，一些评论信息为苹果手机比小米手机好用，通过语法匹配可以判断两句话含义不同，可以分组分别推荐给用户，进一步可以统计每组评论信息的数量，使提问用户更加直观的获取哪些评论支持人数更多。Specifically, groups of comment information with opposite or relative meanings may be selected and recommended to users through grammatical matching. For grammatical matching, refer to the foregoing embodiments. For example, the user question is which is better to use Xiaomi mobile phone or Apple mobile phone, some comment information is that Xiaomi mobile phone is better to use than Apple mobile phone, and some comment information is that Apple mobile phone is better than Xiaomi mobile phone, through grammar matching, it can be judged that the two sentences have different meanings, It can be recommended to users in groups, and further, the number of comments in each group can be counted, so that users who ask questions can more intuitively obtain which comments support more people.
还可以通过语义分析将评论信息进行分组，例如采用前述实施例的方法，确定的内容相似度高于阈值的多条评论信息，将它们划分为一组。又例如，语义分析可以对评论信息的情感或态度进行分析。可以分析评论信息为正面或者负面情感等，可以采用现有的情感分析方法，在此不再赘述，根据评论信息的情感或态度进行分组，分别推荐给用户。The comment information can also be grouped through semantic analysis, for example, by using the method of the aforementioned embodiment, multiple pieces of comment information whose content similarity is determined to be higher than a threshold are classified into a group. For another example, semantic analysis can analyze the emotion or attitude of comment information. The comment information can be analyzed as positive or negative sentiment, etc., and the existing sentiment analysis method can be used, which will not be repeated here, and the comment information is grouped according to the sentiment or attitude of the comment information, and recommended to users respectively.
上述实施例的方法，可以结合用户主动回答提问的机制应用，例如，根据与对象关联(例如购买商品)的用户的信用等级、关联时间、历史回答效率、注册信息、答案采用率中的至少一项信息，选取用户作为回答者，将答案推送至该用户进行回答，同时应用上述实施例的方法向提问用户推送评论信息作为答案。具体的，选取回答者时，用户的信用等级越高，则被选为回答者的概率越高；用户与对象关联时间距提问时间越近，则被选为回答者的概率越高；用户历史回答效率越高，例如，从提问到回答问题的平均时间越短，则被选为回答者的概率越高；用户注册时间越长，则被选为回答者的概率越高；答案采用率则被选为回答者的概率越高。可以将用户的信用等级、关联时间、历史回答效率、注册信息、答案采用率这几项信息分别设置不同的对应的权重，根据加权后的信息确定回答者。The method of the above-mentioned embodiment can be applied in conjunction with the mechanism of the user actively answering questions, for example, according to at least one of the user's credit rating, association time, historical answer efficiency, registration information, and answer adoption rate associated with the object (such as purchasing a product) Item information, select the user as the answerer, push the answer to the user for answering, and apply the method of the above embodiment to push the comment information to the questioning user as the answer. Specifically, when selecting an answerer, the higher the user’s credit rating, the higher the probability of being selected as the answerer; the closer the user’s association time with the object is to the questioning time, the higher the probability of being selected as the answerer; the user’s history The higher the efficiency of answering, for example, the shorter the average time from asking questions to answering questions, the higher the probability of being selected as an answerer; the longer the user registration time, the higher the probability of being selected as an answerer; the answer adoption rate is The higher the probability of being selected as a respondent. Different corresponding weights can be set for the user's credit rating, association time, historical answer efficiency, registration information, and answer adoption rate, and the answerers can be determined according to the weighted information.
上述实施例的方法，针对用户关于对象的提问信息，将该对象的评论信息与提问信息进行匹配，根据匹配结果选取至少一条评论信息作为提问的答案推荐给用户。由于评论信息中包含其他用户关于该对象的评价，利用评论信息可以及时高效地为提问者推荐与问题相关的答案，提高了商品问答中答复的效率，提升用户体验。In the method of the above-mentioned embodiment, according to the question information of the user about the object, the comment information of the object is matched with the question information, and at least one piece of comment information is selected as the answer to the question according to the matching result and recommended to the user. Since the comment information contains other users' evaluations about the object, the comment information can be used to recommend answers related to the question for the questioner in a timely and efficient manner, which improves the efficiency of answering product questions and answers and improves user experience.
下面结合图2描述本公开数据处理方法的另一些实施例。Other embodiments of the data processing method of the present disclosure are described below with reference to FIG. 2 .
图2为本公开数据处理方法另一些实施例的流程图。如图2所示，该实施例的方法包括：步骤S202～S214。Fig. 2 is a flow chart of another embodiment of the data processing method of the present disclosure. As shown in FIG. 2, the method of this embodiment includes: steps S202-S214.
在步骤S202，获取评论信息。In step S202, comment information is obtained.
在步骤S204，建立评论信息与对象信息、用户信息和时间信息的对应关系并存储。In step S204, the corresponding relationship between comment information and object information, user information and time information is established and stored.
对象信息可以包括：对象标识、对象品类信息、对象品牌信息等。根据用户发表评论信息时登录信息可以获取用户信息，可以包括用户个人信息和用户行为信息等，用户个人信息例如，用户标识、用户注册时间、用户信用等级等，用户行为信息例如评论记录、回答提问记录等。用户信息可以通过用户发表评论信息时登录信息(例如用户标识)从系统中查找获取。时间信息包括：评论时间。The object information may include: object identification, object category information, object brand information, and the like. User information can be obtained according to the login information when the user posts comment information, including user personal information and user behavior information. User personal information such as user ID, user registration time, user credit rating, etc. User behavior information such as comment records, answering questions records etc. User information can be retrieved from the system through login information (such as user ID) when the user posts comment information. Time information includes: comment time.
对评论信息进行自动审核保留正常合法的评论信息之后，可以直接建立评论信息与对象信息、用户信息和时间信息的对应关系并存储。也可以参考前述实施例中，对评论信息进行分词、提取关键词、语法信息、语义信息等处理之后，与对象信息、用户信息和时间信息的对应关系并存储。提前对评论信息进行上述处理，可以提高推荐评论信息作为答案的效率。After the comment information is automatically reviewed and the normal and legal comment information is retained, the corresponding relationship between the comment information and the object information, user information and time information can be directly established and stored. You can also refer to the foregoing embodiments, after performing word segmentation, extracting keywords, grammatical information, semantic information, etc. on the comment information, the corresponding relationship with the object information, user information, and time information is stored. Performing the above processing on the comment information in advance can improve the efficiency of recommending the comment information as an answer.
在步骤S206，获取用户关于对象的提问信息。In step S206, the user's question information about the object is acquired.
在步骤S208，查找已存储的对象的评论信息。In step S208, the stored comment information of the object is searched.
在步骤S210，根据评论信息和提问信息的关键词、语法信息和语义信息中的至少一项信息，将评论信息与提问信息进行匹配。In step S210, the comment information is matched with the question information according to at least one item of information among keywords, grammatical information, and semantic information of the comment information and the question information.
步骤S206～S210可以参考前述图1对应的实施例的描述。For steps S206-S210, reference may be made to the description of the embodiment corresponding to FIG. 1 above.
在步骤S212，根据各条评论信息对应的用户信用等级、用户注册信息、评论时间信息和采用率中的至少一项信息，对匹配结果进行修正。In step S212, the matching result is corrected according to at least one piece of information among user credit level, user registration information, comment time information and adoption rate corresponding to each piece of comment information.
在一些实施例中，可以针对不同的用户信用等级设置不同的用户信用等级权重，用户信用等级越高，用户信用等级权重越高。可以针对用户注册时间设置不同的用户注册权重，用户注册时间越早，注册权重越高。可以针对评论时间设置不同的评论时间权重，评论时间与提问信息对应的时间差距越小，评论时间权重越高。可以针对不同的答案采用率设置答案采用率权重，答案采用率越高答案，采用率权重越高。还可以针对用户的评论记录设置评论效果权重，有效评论记录越多，评论效果权重越大。可以针对实际需求选取不同的信息设置不同的权重对匹配结果进行修正。In some embodiments, different user credit level weights may be set for different user credit levels. The higher the user credit level, the higher the user credit level weight. Different user registration weights can be set according to the user registration time. The earlier the user registration time, the higher the registration weight. Different comment time weights can be set for comment time. The smaller the time difference between comment time and question information, the higher the comment time weight. The answer adoption rate weight can be set according to different answer adoption rates. The higher the answer adoption rate is, the higher the adoption rate weight is. The comment effect weight can also be set for the user's comment records, the more valid comment records, the greater the comment effect weight. Different information can be selected according to actual needs and different weights can be set to modify the matching results.
进一步，可以将各项权重进行归一化，将评论信息对应的用户信用等级权重、用户注册权重、评论时间权重、评论效果权重和答案采用率权重中的至少一项与该评论信息对应的匹配度相乘，得到的乘积作为修正后的匹配结果。Further, each weight can be normalized, and at least one of the user credit rating weight, user registration weight, comment time weight, comment effect weight and answer adoption rate weight corresponding to the comment information is matched with the corresponding weight of the comment information. Degrees are multiplied together, and the resulting product is used as the corrected matching result.
在步骤S214，根据修正后的匹配结果向用户推荐至少一条评论信息，作为用户提问的答案。In step S214, at least one piece of comment information is recommended to the user according to the corrected matching result as an answer to the user's question.
可以根据修正后的匹配结果对评论信息进行排序，选取匹配度大于阈值的评论信息推荐给提问用户。具体本步骤可以参考前述图1对应的实施例的描述。The comment information can be sorted according to the corrected matching results, and the comment information whose matching degree is greater than the threshold is selected and recommended to the questioning user. For this step, reference may be made to the description of the embodiment corresponding to FIG. 1 above.
通过上述实施例的方法，可以将更具有参考价值的评论信息推荐给提问用户，进一步提升用户体验。Through the methods of the foregoing embodiments, comment information with more reference value can be recommended to the questioning user, further improving user experience.
本公开还提供一种数据处理装置，下面结合图3进行描述。The present disclosure also provides a data processing device, which will be described below with reference to FIG. 3 .
图3为本公开数据处理装置的一些实施例的结构图。如图3所示，该实施例的装置30包括：提问信息获取模块302，评论信息查找模块304，匹配模块306，答案推荐模块308。Fig. 3 is a structural diagram of some embodiments of a data processing device of the present disclosure. As shown in FIG. 3 , the device 30 of this embodiment includes: a question information acquisition module 302 , a comment information search module 304 , a matching module 306 , and an answer recommendation module 308 .
提问信息获取模块302，用于获取用户关于对象的提问信息。The question information obtaining module 302 is configured to obtain the user's question information about the object.
评论信息查找模块304，用于查找已存储的对象的评论信息。The comment information searching module 304 is configured to search the stored comment information of the object.
评论信息包括从对象评论页面、问答页面、社区类页面和客服系统中至少一处获取的评论信息。The comment information includes comment information obtained from at least one of object comment pages, question-and-answer pages, community pages, and customer service systems.
匹配模块306，用于根据评论信息和提问信息的关键词、语法信息和语义信息中的至少一项信息，将评论信息与提问信息进行匹配。The matching module 306 is configured to match the comment information with the question information according to at least one item of information among keywords, grammatical information, and semantic information of the comment information and the question information.
在一些实施例中，匹配模块306用于对提问信息和评论信息分别进行分词，根据各个词语的词频和包含该词语的训练样本数，分别提取提问信息和评论信息的关键词，根据各条评论信息与提问信息的关键词的相似度，确定各条评论信息与提问信息的关键词匹配度。In some embodiments, the matching module 306 is used to segment the question information and the comment information respectively, extract the keywords of the question information and the comment information according to the word frequency of each word and the number of training samples containing the word, and extract the keywords of the question information and comment information according to each comment The similarity between the keywords of the information and the question information determines the matching degree of keywords between each piece of comment information and the question information.
在一些实施例中，匹配模块306用于对提问信息和评论信息分别进行分词，根据提问信息和评论信息中各个词语的词性，分别确定提问信息和评论信息中各个句子的语法结构，根据提问信息和评论信息中各个句子的语法结构的相似度，确定各条评论信息与提问信息的语法匹配度。In some embodiments, the matching module 306 is used to segment the question information and comment information respectively, determine the grammatical structure of each sentence in the question information and comment information according to the part of speech of each word in the question information and comment information, and determine the grammatical structure of each sentence in the question information and comment information according to the and the similarity of the grammatical structure of each sentence in the comment information, and determine the grammatical matching degree of each comment information and the question information.
在一些实施例中，匹配模块306用于对提问信息和评论信息分别进行分词，分别确定提问信息和评论信息中各个词语的词向量，根据提问信息和评论信息中各个词语的词向量，计算各条评论信息与提问信息的语义匹配度。In some embodiments, the matching module 306 is used to separately segment the question information and comment information, respectively determine the word vectors of each word in the question information and comment information, and calculate the word vectors of each word in the question information and comment information. The semantic matching degree of comment information and question information.
在一些实施例中，匹配模块306用于分别确定各条评论信息与提问信息的关键词匹配度、语法匹配度和语义匹配度，将同一条评论信息对应的关键词匹配度、语法匹配度和语义匹配度进行加权，作为该条评论信息与提问信息的匹配度。In some embodiments, the matching module 306 is used to respectively determine the keyword matching degree, grammatical matching degree and semantic matching degree of each piece of comment information and question information, and compare the keyword matching degree, grammatical matching degree and The semantic matching degree is weighted as the matching degree between the comment information and the question information.
答案推荐模块308，用于根据匹配结果向用户推荐至少一条评论信息，作为用户提问的答案。An answer recommending module 308, configured to recommend at least one piece of comment information to the user according to the matching result as an answer to the user's question.
在一些实施例中，答案推荐模块308用于根据各条评论信息对应的用户信用等级、用户注册信息、评论时间信息和采用率中的至少一项信息，对匹配结果进行修正，根据修正后的匹配结果向用户推荐至少一条评论信息。In some embodiments, the answer recommendation module 308 is used to modify the matching result according to at least one item of information corresponding to each piece of comment information, including user credit rating, user registration information, comment time information and adoption rate, and according to the revised The matching result recommends at least one piece of comment information to the user.
进一步，匹配结果包括各条评论信息与提问信息的匹配度；答案推荐模块308用于将评论信息对应的用户信用等级权重、用户注册权重、评论时间权重和答案采用率权重中的至少一项与该评论信息对应的匹配度相乘，得到的乘积作为修正后的匹配结果；用户信用等级越高，用户信用等级权重越高；用户注册时间越早，注册权重越高；评论时间与提问信息对应的时间差距越小，评论时间权重越高。Further, the matching result includes the matching degree of each piece of comment information and the question information; the answer recommendation module 308 is used to match at least one of the user credit rating weight, user registration weight, comment time weight and answer adoption rate weight corresponding to the comment information with The matching degree corresponding to the comment information is multiplied, and the obtained product is used as the corrected matching result; the higher the user credit rating, the higher the user credit rating weight; the earlier the user registration time, the higher the registration weight; the comment time corresponds to the question information The smaller the time gap of , the higher the weight of comment time.
如图3所示，该数据处理装置30在一些实施例中还可以包括：评论信息处理模块310，用于获取评论信息，建立评论信息与对象信息、用户信息和时间信息的对应关系并存储。As shown in FIG. 3 , in some embodiments, the data processing device 30 may further include: a comment information processing module 310 , configured to acquire comment information, establish and store correspondence between comment information and object information, user information, and time information.
本公开的实施例中的数据处理装置可各由各种计算设备或计算机系统来实现，下面结合图4以及图5进行描述。The data processing devices in the embodiments of the present disclosure may be implemented by various computing devices or computer systems, which will be described below in conjunction with FIG. 4 and FIG. 5 .
图4为本公开数据处理装置的一些实施例的结构图。如图4所示，该实施例的装置40包括：存储器410以及耦接至该存储器410的处理器420，处理器420被配置为基于存储在存储器410中的指令，执行本公开中任意一些实施例中的数据处理方法。Fig. 4 is a structural diagram of some embodiments of the data processing device of the present disclosure. As shown in FIG. 4 , the apparatus 40 of this embodiment includes: a memory 410 and a processor 420 coupled to the memory 410 , the processor 420 is configured to execute any implementation in the present disclosure based on instructions stored in the memory 410 . The data processing method in the example.
其中，存储器410例如可以包括系统存储器、固定非易失性存储介质等。系统存储器例如存储有操作系统、应用程序、引导装载程序(Boot Loader)、数据库以及其他程序等。Wherein, the memory 410 may include, for example, a system memory, a fixed non-volatile storage medium, and the like. The system memory stores, for example, an operating system, an application program, a boot loader (Boot Loader), a database, and other programs.
图5为本公开数据处理装置的另一些实施例的结构图。如图5所示，该实施例的装置50包括：存储器510以及处理器520，分别与存储器410以及处理器420类似。还可以包括输入输出接口530、网络接口540、存储接口550等。这些接口530，540，550以及存储器510和处理器520之间例如可以通过总线560连接。其中，输入输出接口530为显示器、鼠标、键盘、触摸屏等输入输出设备提供连接接口。网络接口540为各种联网设备提供连接接口，例如可以连接到数据库服务器或者云端存储服务器等。存储接口550为SD卡、U盘等外置存储设备提供连接接口。Fig. 5 is a structural diagram of other embodiments of the data processing device of the present disclosure. As shown in FIG. 5 , the device 50 of this embodiment includes: a memory 510 and a processor 520 , which are similar to the memory 410 and the processor 420 respectively. It may also include an input/output interface 530, a network interface 540, a storage interface 550, and the like. These interfaces 530 , 540 , 550 as well as the memory 510 and the processor 520 may be connected via a bus 560 , for example. Wherein, the input and output interface 530 provides a connection interface for input and output devices such as a display, a mouse, a keyboard, and a touch screen. The network interface 540 provides connection interfaces for various networking devices, for example, it can be connected to a database server or a cloud storage server. The storage interface 550 provides connection interfaces for external storage devices such as SD cards and U disks.
本领域内的技术人员应当明白，本公开的实施例可提供为方法、系统、或计算机程序产品。因此，本公开可采用完全硬件实施例、完全软件实施例、或结合软件和硬件方面的实施例的形式。而且，本公开可采用在一个或多个其中包含有计算机可用程序代码的计算机可用非瞬时性存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。Those skilled in the art should understand that the embodiments of the present disclosure may be provided as methods, systems, or computer program products. Accordingly, the present disclosure can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present disclosure may take the form of a computer program product embodied on one or more computer-usable non-transitory storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein. .
本公开是参照根据本公开实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解为可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器，使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。The present disclosure is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present disclosure. It should be understood that each process and/or block in the flowchart and/or block diagram, and a combination of processes and/or blocks in the flowchart and/or block diagram can be realized by computer program instructions. These computer program instructions may be provided to a general purpose computer, special purpose computer, embedded processor, or processor of other programmable data processing equipment to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing equipment produce a An apparatus for realizing the functions specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中，使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品，该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to operate in a specific manner, such that the instructions stored in the computer-readable memory produce an article of manufacture comprising instruction means, the instructions The device realizes the function specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上，使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理，从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。These computer program instructions can also be loaded onto a computer or other programmable data processing device, causing a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process, thereby The instructions provide steps for implementing the functions specified in the flow chart or blocks of the flowchart and/or the block or blocks of the block diagrams.
以上所述仅为本公开的较佳实施例，并不用以限制本公开，凡在本公开的精神和原则之内，所作的任何修改、等同替换、改进等，均应包含在本公开的保护范围之内。The above descriptions are only preferred embodiments of the present disclosure, and are not intended to limit the present disclosure. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present disclosure shall be included in the protection of the present disclosure. within range.
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810308511.8ACN110362662A (en) | 2018-04-09 | 2018-04-09 | Data processing method, device and computer readable storage medium |
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810308511.8ACN110362662A (en) | 2018-04-09 | 2018-04-09 | Data processing method, device and computer readable storage medium |
| Publication Number | Publication Date |
|---|---|
| CN110362662Atrue CN110362662A (en) | 2019-10-22 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810308511.8APendingCN110362662A (en) | 2018-04-09 | 2018-04-09 | Data processing method, device and computer readable storage medium |
| Country | Link |
|---|---|
| CN (1) | CN110362662A (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111797209A (en)* | 2020-02-26 | 2020-10-20 | 北京沃东天骏信息技术有限公司 | Customer service robot response method and device |
| CN112801745A (en)* | 2021-02-02 | 2021-05-14 | 李海涛 | Big data platform based online comment validity recommendation method |
| CN112989020A (en)* | 2019-12-17 | 2021-06-18 | 北京沃东天骏信息技术有限公司 | Information processing method, apparatus and computer readable storage medium |
| CN113051380A (en)* | 2021-03-23 | 2021-06-29 | 北京百度网讯科技有限公司 | Information generation method and device, electronic equipment and storage medium |
| CN113610247A (en)* | 2021-07-22 | 2021-11-05 | 北京中交兴路信息科技有限公司 | Fault help seeking method and device for freight vehicle, storage medium and terminal |
| CN116226677A (en)* | 2023-05-09 | 2023-06-06 | 北京搜狐新媒体信息技术有限公司 | Parallel corpus construction method and device, storage medium and electronic equipment |
| CN118917917A (en)* | 2024-07-25 | 2024-11-08 | 北京达佳互联信息技术有限公司 | Information display method and device, electronic equipment and storage medium |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102279894A (en)* | 2011-09-19 | 2011-12-14 | 嘉兴亿言堂信息科技有限公司 | Method for searching, integrating and providing comment information based on semantics and searching system |
| CN104536980A (en)* | 2014-12-05 | 2015-04-22 | 百度在线网络技术（北京）有限公司 | To-be-commented item quality information determination method and device |
| CN105701253A (en)* | 2016-03-04 | 2016-06-22 | 南京大学 | Chinese natural language interrogative sentence semantization knowledge base automatic question-answering method |
| CN106709007A (en)* | 2016-12-23 | 2017-05-24 | 北京奇虎科技有限公司 | Automobile search result display method and automobile search result display device |
| CN106997376A (en)* | 2017-02-28 | 2017-08-01 | 浙江大学 | The problem of one kind is based on multi-stage characteristics and answer sentence similarity calculating method |
| CN107833088A (en)* | 2017-10-17 | 2018-03-23 | 北京百度网讯科技有限公司 | Content providing, device and smart machine |
| CN107844533A (en)* | 2017-10-19 | 2018-03-27 | 云南大学 | A kind of intelligent Answer System and analysis method |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102279894A (en)* | 2011-09-19 | 2011-12-14 | 嘉兴亿言堂信息科技有限公司 | Method for searching, integrating and providing comment information based on semantics and searching system |
| CN104536980A (en)* | 2014-12-05 | 2015-04-22 | 百度在线网络技术（北京）有限公司 | To-be-commented item quality information determination method and device |
| CN105701253A (en)* | 2016-03-04 | 2016-06-22 | 南京大学 | Chinese natural language interrogative sentence semantization knowledge base automatic question-answering method |
| CN106709007A (en)* | 2016-12-23 | 2017-05-24 | 北京奇虎科技有限公司 | Automobile search result display method and automobile search result display device |
| CN106997376A (en)* | 2017-02-28 | 2017-08-01 | 浙江大学 | The problem of one kind is based on multi-stage characteristics and answer sentence similarity calculating method |
| CN107833088A (en)* | 2017-10-17 | 2018-03-23 | 北京百度网讯科技有限公司 | Content providing, device and smart machine |
| CN107844533A (en)* | 2017-10-19 | 2018-03-27 | 云南大学 | A kind of intelligent Answer System and analysis method |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112989020A (en)* | 2019-12-17 | 2021-06-18 | 北京沃东天骏信息技术有限公司 | Information processing method, apparatus and computer readable storage medium |
| CN111797209A (en)* | 2020-02-26 | 2020-10-20 | 北京沃东天骏信息技术有限公司 | Customer service robot response method and device |
| CN112801745A (en)* | 2021-02-02 | 2021-05-14 | 李海涛 | Big data platform based online comment validity recommendation method |
| CN113051380A (en)* | 2021-03-23 | 2021-06-29 | 北京百度网讯科技有限公司 | Information generation method and device, electronic equipment and storage medium |
| CN113051380B (en)* | 2021-03-23 | 2023-07-25 | 北京百度网讯科技有限公司 | Information generation method, device, electronic equipment and storage medium |
| CN113610247A (en)* | 2021-07-22 | 2021-11-05 | 北京中交兴路信息科技有限公司 | Fault help seeking method and device for freight vehicle, storage medium and terminal |
| CN116226677A (en)* | 2023-05-09 | 2023-06-06 | 北京搜狐新媒体信息技术有限公司 | Parallel corpus construction method and device, storage medium and electronic equipment |
| CN116226677B (en)* | 2023-05-09 | 2023-07-14 | 北京搜狐新媒体信息技术有限公司 | Parallel corpus construction method and device, storage medium and electronic equipment |
| CN118917917A (en)* | 2024-07-25 | 2024-11-08 | 北京达佳互联信息技术有限公司 | Information display method and device, electronic equipment and storage medium |
| Publication | Publication Date | Title |
|---|---|---|
| US10748164B2 (en) | Analyzing sentiment in product reviews | |
| US11334635B2 (en) | Domain specific natural language understanding of customer intent in self-help | |
| US11907274B2 (en) | Hyper-graph learner for natural language comprehension | |
| US10719665B2 (en) | Unsupervised neural based hybrid model for sentiment analysis of web/mobile application using public data sources | |
| CN105989040B (en) | Intelligent question and answer method, device and system | |
| CN107391493B (en) | Public opinion information extraction method and device, terminal equipment and storage medium | |
| CN110263248B (en) | Information pushing method, device, storage medium and server | |
| CN110362662A (en) | Data processing method, device and computer readable storage medium | |
| US10223442B2 (en) | Prioritizing survey text responses | |
| CN111666761B (en) | Fine-grained sentiment analysis model training method and device | |
| CN108491377A (en) | A kind of electric business product comprehensive score method based on multi-dimension information fusion | |
| CN110516033B (en) | Method and device for calculating user preference | |
| EP2778969A1 (en) | Search result ranking using query clustering | |
| CN110706028A (en) | Commodity evaluation emotion analysis system based on attribute characteristics | |
| WO2019242453A1 (en) | Information processing method and device, storage medium, and electronic device | |
| CN113821588A (en) | Text processing method and device, electronic equipment and storage medium | |
| CN118350894A (en) | Commodity recommendation method and device and electronic equipment | |
| CN112132368A (en) | Information processing method and device, computing equipment and storage medium | |
| CN113722487A (en) | User emotion analysis method, device and equipment and storage medium | |
| TWI712948B (en) | Method for document sentimental analysis, apparatus and computer program product thereof | |
| CN116860969A (en) | A customer review analysis method, system, equipment and medium | |
| US20140324524A1 (en) | Evolving a capped customer linkage model using genetic models | |
| CN115080741A (en) | Questionnaire survey analysis method, device, storage medium and equipment | |
| CN114329206A (en) | Title generation method and device, electronic equipment and computer readable medium | |
| CN112463959A (en) | Service processing method based on uplink short message and related equipment |
| 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 | ||
| RJ01 | Rejection of invention patent application after publication | Application publication date:20191022 | |
| RJ01 | Rejection of invention patent application after publication |