WO2021151358A1 - Triage information recommendation method and apparatus based on interpretation model, and device and medium - Google Patents

Abstract

The present application relates to the technical field of artificial intelligence, and is applied to the field of smart medical treatment, so as to promote the construction of smart cities. Disclosed are a triage information recommendation method and apparatus based on an interpretation model, and a device and a medium. The method comprises: inputting acquired feature information text, which is input by an object to be subjected to triage, into a preset interpretation model, and performing interpretation processing on the feature information text to obtain an interpretation result corresponding to the feature information text; comparing a matching value corresponding to each word in the feature information text with a preset matching threshold value to determine whether there are triage interpretation words in the feature information text; when there are triage interpretation words in the feature information text, inputting the feature information text into a preset triage decision model to perform triage processing on the feature information text to obtain triage information; and associatively pushing the triage information and the triage interpretation words to said object. According to the present application, the utilization rate and efficiency of a preset triage decision model are improved, and the accuracy rate of triage information is improved.

Description

基于解释模型的分诊信息推荐方法、装置、设备及介质Method, device, equipment and medium for recommending triage information based on interpretation model

本申请要求于2020年9月8日提交中国专利局、申请号为202010935273.0，发明名称为“基于解释模型的分诊信息推荐方法、装置、设备及介质”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。This application claims the priority of the Chinese patent application filed with the Chinese Patent Office on September 8, 2020, the application number is 202010935273.0, and the invention title is "Interpretation Model-based Triage Information Recommendation Method, Apparatus, Equipment, and Medium", all of which The content is incorporated in this application by reference.

技术领域Technical field

本申请涉及人工智能技术领域，尤其涉及一种基于解释模型的分诊信息推荐方法、装置、设备及介质。This application relates to the field of artificial intelligence technology, and in particular to a method, device, equipment, and medium for recommending triage information based on an interpretation model.

 To

背景技术Background technique

随着科学技术的发展，人工智能技术也随之发展。例如，在医疗行业中的分诊***，分诊***的应用可以替代分诊台的医护人员进行分诊推荐，有效缓解医护人员的压力。With the development of science and technology, artificial intelligence technology has also developed. For example, in the triage system in the medical industry, the application of the triage system can replace the medical staff at the triage table to make triage recommendations, effectively alleviating the pressure on the medical staff.

发明人意识到，现有技术中的分诊***，需要患者录入分诊文本至分诊***之后，分诊***根据分诊文本推送分诊信息，该方案的不足之处在于：在患者在不小心甚至恶意输入错误文本（与病情和分诊无关的文本）时，分诊***依旧会根据输入的错误文本输出分诊信息，在该情况下，分诊***的利用率较低；并且，患者通常是通过口语化的描述录入分诊文本，因此，分诊***对分诊文本的识别往往较为片面，如此会导致最终输出的分诊信息准确率较低。The inventor realizes that the triage system in the prior art requires the patient to input the triage text to the triage system, and the triage system pushes the triage information according to the triage text. The disadvantage of this solution is that: Be careful and even maliciously input the wrong text (text not related to the condition and triage), the triage system will still output triage information according to the input wrong text. In this case, the utilization rate of the triage system is low; moreover, the patient The triage text is usually entered through colloquial description. Therefore, the recognition of triage text by the triage system is often one-sided, which will result in a lower accuracy of the final output of the triage information.

 To

申请内容Application content

本申请实施例提供一种基于解释模型的分诊信息推荐方法、装置、设备及介质，以解决分诊***利用率较低以及分诊信息准确率较低的问题。The embodiments of the present application provide a method, device, equipment, and medium for recommending triage information based on an interpretation model to solve the problems of low utilization of the triage system and low accuracy of triage information.

一种基于解释模型的分诊信息推荐方法，包括：A method for recommending triage information based on an interpretation model, including:

获取待分诊对象录入的特征信息文本；Obtain the characteristic information text entered by the subject to be triaged;

将所述特征信息文本输入至预设解释模型，对所述特征信息文本进行解释处理，得到与所述特征信息文本对应的解释结果；所述解释结果中包含预设分诊参考向量与所述特征信息文本中各词语对应的匹配值；The feature information text is input into a preset interpretation model, and the feature information text is interpreted to obtain an interpretation result corresponding to the feature information text; the interpretation result includes the preset triage reference vector and the The matching value corresponding to each word in the feature information text;

比对所述特征信息文本的各词语对应的匹配值以及预设匹配阈值，确定所述特征信息文本中是否存在分诊解释词语；所述分诊解释词语是指所述特征信息文本中与大于或等于所述预设匹配阈值的匹配值对应的词语；Compare the matching value corresponding to each word of the characteristic information text and the preset matching threshold to determine whether there are triage interpretation words in the characteristic information text; Or a word corresponding to a matching value equal to the preset matching threshold;

在所述特征信息文本中存在分诊解释词语时，将所述特征信息文本输入至预设分诊决策模型中，对所述特征信息文本进行分诊处理，得到与所述特征信息文本对应的分诊信息；When there are triage interpretation words in the feature information text, input the feature information text into a preset triage decision model, and perform triage processing on the feature information text to obtain the corresponding feature information text Triage information;

将所述分诊信息以及所述分诊解释词语关联推送至所述待分诊对象。Push the triage information and the triage explanation word association to the object to be triaged.

一种基于解释模型的分诊信息推荐装置，包括：An interpretation model-based triage information recommendation device, including:

文本获取模块，用于获取待分诊对象录入的特征信息文本；The text acquisition module is used to acquire the characteristic information text entered by the object to be triaged;

解释处理模块，用于将所述特征信息文本输入至预设解释模型，对所述特征信息文本进行解释处理，得到与所述特征信息文本对应的解释结果；所述解释结果中包含预设分诊参考向量与所述特征信息文本中各词语对应的匹配值；The interpretation processing module is used to input the feature information text into a preset interpretation model, perform interpretation processing on the feature information text, and obtain an interpretation result corresponding to the feature information text; the interpretation result includes the preset score The matching value between the diagnosis reference vector and each word in the characteristic information text;

匹配比对模块，用于比对所述特征信息文本的各词语对应的匹配值以及预设匹配阈值，确定所述特征信息文本中是否存在分诊解释词语；所述分诊解释词语是指所述特征信息文本中与大于或等于所述预设匹配阈值的匹配值对应的词语；The matching and comparison module is used to compare the matching value corresponding to each word of the characteristic information text and the preset matching threshold to determine whether there are triage interpretation words in the characteristic information text; the triage interpretation words refer to all Words in the feature information text that correspond to a matching value greater than or equal to the preset matching threshold;

文本分类模块，用于在所述特征信息文本中存在分诊解释词语时，将所述特征信息文本输入至预设分诊决策模型中，对所述特征信息文本进行分诊处理，得到与所述特征信息文本对应的分诊信息；The text classification module is used to input the feature information text into the preset triage decision model when there are triage interpretation words in the feature information text, and perform triage processing on the feature information text to obtain the The triage information corresponding to the text of the characteristic information;

分诊信息推送模块，用于将所述分诊信息以及所述分诊解释词语关联推送至所述待分诊对象。The triage information push module is configured to push the triage information and the triage explanation word association to the object to be triaged.

一种计算机设备，包括存储器、处理器以及存储在所述存储器中并可在所述处理器上运行的计算机可读指令，所述处理器执行所述计算机可读指令时实现如下步骤：A computer device includes a memory, a processor, and computer-readable instructions that are stored in the memory and can run on the processor, and the processor implements the following steps when the processor executes the computer-readable instructions:

获取待分诊对象录入的特征信息文本；Obtain the characteristic information text entered by the subject to be triaged;

将所述特征信息文本输入至预设解释模型，对所述特征信息文本进行解释处理，得到与所述特征信息文本对应的解释结果；所述解释结果中包含预设分诊参考向量与所述特征信息文本中各词语对应的匹配值；The feature information text is input into a preset interpretation model, and the feature information text is interpreted to obtain an interpretation result corresponding to the feature information text; the interpretation result includes the preset triage reference vector and the The matching value corresponding to each word in the feature information text;

比对所述特征信息文本的各词语对应的匹配值以及预设匹配阈值，确定所述特征信息文本中是否存在分诊解释词语；所述分诊解释词语是指所述特征信息文本中与大于或等于所述预设匹配阈值的匹配值对应的词语；Compare the matching value corresponding to each word of the characteristic information text and the preset matching threshold to determine whether there are triage interpretation words in the characteristic information text; Or a word corresponding to a matching value equal to the preset matching threshold;

在所述特征信息文本中存在分诊解释词语时，将所述特征信息文本输入至预设分诊决策模型中，对所述特征信息文本进行分诊处理，得到与所述特征信息文本对应的分诊信息；When there are triage interpretation words in the feature information text, input the feature information text into a preset triage decision model, and perform triage processing on the feature information text to obtain the corresponding feature information text Triage information;

将所述分诊信息以及所述分诊解释词语关联推送至所述待分诊对象。Push the triage information and the triage explanation word association to the object to be triaged.

一个或多个存储有计算机可读指令的可读存储介质，所述计算机可读指令被一个或多个处理器执行时，使得所述一个或多个处理器执行如下步骤：One or more readable storage media storing computer readable instructions, when the computer readable instructions are executed by one or more processors, the one or more processors execute the following steps:

获取待分诊对象录入的特征信息文本；Obtain the characteristic information text entered by the subject to be triaged;

将所述特征信息文本输入至预设解释模型，对所述特征信息文本进行解释处理，得到与所述特征信息文本对应的解释结果；所述解释结果中包含预设分诊参考向量与所述特征信息文本中各词语对应的匹配值；The feature information text is input into a preset interpretation model, and the feature information text is interpreted to obtain an interpretation result corresponding to the feature information text; the interpretation result includes the preset triage reference vector and the The matching value corresponding to each word in the feature information text;

比对所述特征信息文本的各词语对应的匹配值以及预设匹配阈值，确定所述特征信息文本中是否存在分诊解释词语；所述分诊解释词语是指所述特征信息文本中与大于或等于所述预设匹配阈值的匹配值对应的词语；Compare the matching value corresponding to each word of the characteristic information text and the preset matching threshold to determine whether there are triage interpretation words in the characteristic information text; Or a word corresponding to a matching value equal to the preset matching threshold;

在所述特征信息文本中存在分诊解释词语时，将所述特征信息文本输入至预设分诊决策模型中，对所述特征信息文本进行分诊处理，得到与所述特征信息文本对应的分诊信息；When there are triage interpretation words in the feature information text, input the feature information text into a preset triage decision model, and perform triage processing on the feature information text to obtain the corresponding feature information text Triage information;

将所述分诊信息以及所述分诊解释词语关联推送至所述待分诊对象。上述基于解释模型的分诊信息推荐方法、装置、设备及介质，通过获取待分诊对象录入的特征信息文本；将所述特征信息文本输入至预设解释模型，对所述特征信息文本进行解释处理，得到与所述特征信息文本对应的解释结果；所述解释结果中包含预设分诊参考向量与所述特征信息文本中各词语对应的匹配值；比对所述特征信息文本的各词语对应的匹配值以及预设匹配阈值，确定所述特征信息文本中是否存在分诊解释词语；所述分诊解释词语是指所述特征信息文本中与大于或等于所述预设匹配阈值的匹配值对应的词语；在所述特征信息文本中存在分诊解释词语时，将所述特征信息文本输入至预设分诊决策模型中，对所述特征信息文本进行分诊处理，得到与所述特征信息文本对应的分诊信息；将所述分诊信息以及所述分诊解释词语关联推送至所述待分诊对象。Push the triage information and the triage explanation word association to the object to be triaged. The above-mentioned method, device, device, and medium of triage information recommendation based on the interpretation model obtain the feature information text entered by the object to be triaged; input the feature information text into the preset interpretation model to interpret the feature information text Processing to obtain the interpretation result corresponding to the characteristic information text; the interpretation result includes the matching value of the preset triage reference vector and each word in the characteristic information text; compare each word in the characteristic information text Corresponding matching values and preset matching thresholds to determine whether there are triage interpretation words in the characteristic information text; the triage interpretation words refer to matches in the characteristic information text that are greater than or equal to the preset matching threshold The word corresponding to the value; when there are triage interpretation words in the feature information text, the feature information text is input into the preset triage decision model, and the feature information text is triaged to obtain the The triage information corresponding to the feature information text; and push the triage information and the triage explanation word association to the object to be triaged.

本申请通过引入预设解释模型，在特征信息文本中存在分诊解释词语时，才将特征信息文本输入至预设分诊决策模型进行分诊，提高了预设分诊决策模型的利用率以及效率；并且，本申请中将得到的分诊信息与分诊解释词语关联推送至待分诊对象，令待分诊对象可以了解给出该分诊信息的依据，在保证分诊信息的准确率的同时，提升了用户体验。In this application, by introducing a preset interpretation model, when there are triage interpretation words in the feature information text, the feature information text is input into the preset triage decision model for triage, which improves the utilization rate of the preset triage decision model. Efficiency; In addition, the triage information obtained in this application is associated with the triage explanation words and pushed to the subject to be triaged, so that the subject to be triaged can understand the basis for giving the triage information, and ensure the accuracy of the triage information At the same time, the user experience is improved.

本申请的一个或多个实施例的细节在下面的附图和描述中提出，本申请的其他特征和优点将从说明书、附图以及权利要求变得明显。The details of one or more embodiments of the present application are presented in the following drawings and description, and other features and advantages of the present application will become apparent from the description, drawings and claims.

 To

附图说明Description of the drawings

为了更清楚地说明本申请实施例的技术方案，下面将对本申请实施例的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性的前提下，还可以根据这些附图获得其他的附图。In order to explain the technical solutions of the embodiments of the present application more clearly, the following will briefly introduce the drawings that need to be used in the description of the embodiments of the present application. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative labor.

图1是本申请一实施例中基于解释模型的分诊信息推荐方法的一应用环境示意图；FIG. 1 is a schematic diagram of an application environment of a method for recommending triage information based on an interpretation model in an embodiment of the present application;

图2是本申请一实施例中基于解释模型的分诊信息推荐方法的一流程图；2 is a flowchart of a method for recommending triage information based on an interpretation model in an embodiment of the present application;

图3是本申请一实施例中基于解释模型的分诊信息推荐方法中步骤S20的一流程图；3 is a flowchart of step S20 in the method for recommending triage information based on the interpretation model in an embodiment of the present application;

图4是本申请一实施例中基于解释模型的分诊信息推荐方法中步骤S40的一流程图；4 is a flowchart of step S40 in the method for recommending triage information based on the interpretation model in an embodiment of the present application;

图5是本申请一实施例中基于解释模型的分诊信息推荐装置的一原理框图；Fig. 5 is a functional block diagram of a device for recommending triage information based on an interpretation model in an embodiment of the present application;

图6是本申请一实施例中基于解释模型的分诊信息推荐装置中解释处理模块的一原理框图；6 is a functional block diagram of the interpretation processing module in the triage information recommendation device based on the interpretation model in an embodiment of the present application;

图7是本申请一实施例中基于解释模型的分诊信息推荐装置中文本分类模块的一原理框图；FIG. 7 is a functional block diagram of a text classification module in an interpretation model-based triage information recommendation device in an embodiment of the present application;

图8是本申请一实施例中计算机设备的一示意图。Fig. 8 is a schematic diagram of a computer device in an embodiment of the present application.

 To

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清楚、完整地描述，显然，所描述的实施例是本申请一部分实施例，而不是全部的实施例。基于本申请中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都属于本申请保护的范围。The technical solutions in the embodiments of the present application will be described clearly and completely in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all of them. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative work shall fall within the protection scope of this application.

本申请实施例提供的基于解释模型的分诊信息推荐方法，该基于解释模型的分诊信息推荐方法可应用如图1所示的应用环境中。具体地，该基于解释模型的分诊信息推荐方法应用在基于解释模型的分诊信息推荐***中，该基于解释模型的分诊信息推荐***包括如图1所示的客户端和服务器，客户端与服务器通过网络进行通信，用于解决分诊***利用率较低以及分诊信息准确率较低的问题。其中，客户端又称为用户端，是指与服务器相对应,为客户提供本地服务的程序。客户端可安装在但不限于各种个人计算机、笔记本电脑、智能手机、平板电脑和便携式可穿戴设备上。服务器可以用独立的服务器或者是多个服务器组成的服务器集群来实现。The method for recommending triage information based on the interpretation model provided by the embodiment of the present application can be applied to the application environment as shown in FIG. 1. Specifically, the method for recommending triage information based on the interpretation model is applied in a triage information recommendation system based on the interpretation model. The triage information recommendation system based on the interpretation model includes a client and a server as shown in FIG. Communicate with the server through the network to solve the problems of low utilization of the triage system and low accuracy of triage information. Among them, the client is also called the client, which refers to the program that corresponds to the server and provides local services to the client. The client can be installed on, but not limited to, various personal computers, notebook computers, smart phones, tablet computers, and portable wearable devices. The server can be implemented as an independent server or a server cluster composed of multiple servers.

在一实施例中，如图2所示，提供一种基于解释模型的分诊信息推荐方法，以该方法应用在图1中的服务器为例进行说明，包括如下步骤：In an embodiment, as shown in FIG. 2, a method for recommending triage information based on an interpretation model is provided. Taking the method applied to the server in FIG. 1 as an example for description, the method includes the following steps:

S10S10 ：获取待分诊对象录入的特征信息文本。: Obtain the characteristic information text entered by the object to be triaged.

示例性地，假设该方法应用于线下的医疗行业，则该待分诊对象可以为患者；假设该方法应用于线上的应用程序，则该待分诊对象可以为客户端的用户。特征信息文本包括但不限于待分诊对象的基本信息（如年龄、身高等）、症状信息（如病情信息等）等。Illustratively, assuming that the method is applied to the offline medical industry, the subject to be triaged may be a patient; assuming that the method is applied to an online application, the subject to be triaged may be a user of the client. The characteristic information text includes but is not limited to basic information (such as age, height, etc.), symptom information (such as condition information, etc.) of the subject to be triaged.

S20S20 ：将所述特征信息文本输入至预设解释模型，对所述特征信息文本进行解释处理，得到与所述特征信息文本对应的解释结果；所述解释结果中包含预设分诊参考向量与所述特征信息文本中各词语对应的匹配值。: Input the feature information text into a preset interpretation model, perform interpretation processing on the feature information text, and obtain an interpretation result corresponding to the feature information text; the interpretation result includes the preset triage reference vector and the Describe the matching value corresponding to each word in the feature information text.

其中，预设解释模型是基于循环神经网络模型以及注意力模型构建的模型。解释处理指的是获取特征信息文本中各词语的匹配值的过程。匹配值指的是特征信息文本中各词语对应的实体向量与预设分诊参考向量之间的匹配程度。Among them, the preset interpretation model is a model constructed based on the recurrent neural network model and the attention model. Interpretation processing refers to the process of obtaining the matching value of each word in the characteristic information text. The matching value refers to the matching degree between the entity vector corresponding to each word in the feature information text and the preset triage reference vector.

在一具体实施方式中，如图3所示，步骤S20包括如下步骤：In a specific embodiment, as shown in FIG. 3, step S20 includes the following steps:

S201S201 ：根据预设词语实体库，对所述特征信息文本进行分词处理，得到与所述特征信息文本对应的各字词实体。: Perform word segmentation processing on the characteristic information text according to a preset word entity database to obtain each word entity corresponding to the characteristic information text.

其中，预设词语实体库指的是采用历史数据（如历史药物信息、历史住院信息等）中统计的所有医学实体（如咳嗽、感冒等症状实体，亦或者具体的疾病实体，如白血病等）构建得到的数据库。Among them, the preset word entity database refers to all medical entities (such as cough, cold and other symptom entities, or specific disease entities, such as leukemia, etc.) from historical data (such as historical drug information, historical hospitalization information, etc.) Build the resulting database.

具体地，在获取待分诊对象的特征信息文本之后，将特征信息文本输入至预设解释模型中，根据该预设解释模型中的预设词语实体库，对特征信息文本进行分词处理，得到与特征信息文本对应的各字词实体。Specifically, after acquiring the characteristic information text of the subject to be triaged, the characteristic information text is input into the preset interpretation model, and the characteristic information text is segmented according to the preset word entity database in the preset interpretation model to obtain Each word entity corresponding to the feature information text.

示例性地，假设特征信息文本为“发现咳嗽三天”，在分词处理后得到的各字词实体为“发现”、“咳嗽”以及“三天”。Exemplarily, assuming that the characteristic information text is "found cough for three days", the word entities obtained after word segmentation are "discovered", "cough" and "three days".

S202S202 ：将各字词实体输入至所述预设解释模型的循环神经网络模型中，得到与各所述字词实体对应的实体向量。: Input each word entity into the cyclic neural network model of the preset interpretation model to obtain an entity vector corresponding to each word entity.

作为优选，循环神经网络模型选择LSTM（Long Short-Term Memory，长短期记忆网络）。As a preference, the recurrent neural network model chooses LSTM (Long Short-Term Memory, long and short-term memory network).

具体地，在根据预设词语实体库，对所述特征信息文本进行分词处理，得到与所述特征信息文本对应的各字词实体之后，将各字词实体输入至预设解释模型的循环神经网路模型中，对各字词实体进行正向词向量编码，得到与各字词实体对应的正向词向量；对各字词实体进行反向词向量编码，得到与各字词实体对应的反向词向量；对各词语的正向词向量和反向词向量进行融合表征，得到与各字词实体对应的实体向量，进而可以保证各实体向量均可以融入特征信息文本中的上下文信息。Specifically, after performing word segmentation processing on the characteristic information text according to a preset word entity database to obtain each word entity corresponding to the characteristic information text, input each word entity to the cyclic nerve of the preset interpretation model In the network model, forward word vector encoding is performed on each word entity to obtain the forward word vector corresponding to each word entity; reverse word vector encoding is performed on each word entity to obtain the corresponding word entity Reverse word vector: The forward word vector and reverse word vector of each word are fused and characterized, and the entity vector corresponding to each word entity is obtained, thereby ensuring that each entity vector can be integrated into the context information in the feature information text.

S203S203 ：获取预设分诊参考向量，将各所述实体向量与所述预设分诊参考向量输入至所述预设解释模型中的注意力模型中，确定各所述实体向量与所述预设分诊参考向量之间的匹配值。: Obtain a preset triage reference vector, input each of the entity vectors and the preset triage reference vector into the attention model in the preset interpretation model, and determine each of the entity vectors and the preset Match value between triage reference vectors.

其中，预设分诊参考向量是基于历史医疗数据中的各医疗体系的字词得到的，该预设分诊参考向量用于判定特征信息文本中各实体向量是否符合医疗体系下的实体向量。注意力模型用于确定各实体向量与预设分诊参考向量之间的匹配值，该注意力模型是基于注意力机制构建的。The preset triage reference vector is obtained based on the words of each medical system in the historical medical data, and the preset triage reference vector is used to determine whether each entity vector in the feature information text matches the entity vector under the medical system. The attention model is used to determine the matching value between each entity vector and the preset triage reference vector. The attention model is constructed based on the attention mechanism.

具体地，在将各字词实体输入至所述预设解释模型的循环神经网络模型中，得到与各所述字词实体对应的实体向量之后，获取预设分诊参考向量；将各实体向量以及预设分诊参考向量输入至预设解释模型中的注意力模型中，采用距离度量函数确定各实体向量与参考向量之间的参考距离；通过softmax（逻辑回归）网络对各参考距离进行归一化处理之后，得到各实体向量与预设分诊参考向量之间的匹配值。Specifically, after each word entity is input into the recurrent neural network model of the preset interpretation model, and after the entity vector corresponding to each word entity is obtained, a preset triage reference vector is obtained; and each entity vector And the preset triage reference vector is input into the attention model in the preset interpretation model, and the distance measurement function is used to determine the reference distance between each entity vector and the reference vector; each reference distance is classified through the softmax (logistic regression) network After the unified processing, the matching value between each entity vector and the preset triage reference vector is obtained.

在一具体实施方式中，在步骤S203之前，也即获取预设分诊参考向量之前，包括：In a specific implementation, before step S203, that is, before obtaining the preset triage reference vector, the method includes:

S205S205 ：获取样本数据集，所述样本数据集中包含至少一个样本信息；所述样本信息中包含至少一个样本决策词语。: Obtain a sample data set, the sample data set includes at least one sample information; the sample information includes at least one sample decision word.

其中，样本数据集指的是采集不同的历史医疗数据（如历史住院信息、历史药物信息等）得到的样本集合。样本信息可以为一个历史住院信息、一个历史药物信息等。样本决策词语指的是在历史住院信息中与医疗体系相关的词语（如感冒、咳嗽等各种病情词语）。Among them, the sample data set refers to the collection of samples obtained by collecting different historical medical data (such as historical hospitalization information, historical drug information, etc.). The sample information can be a piece of historical hospitalization information, a piece of historical drug information, and so on. The sample decision words refer to the words related to the medical system in the historical hospitalization information (such as cold, cough and other illness words).

S206S206 ：将各所述样本决策词语输入至预设词向量模型中，对各所述样本决策词语进行向量表示，得到与各所述样本决策词语对应的决策词向量。: Input each of the sample decision words into a preset word vector model, and perform a vector representation on each of the sample decision words to obtain a decision word vector corresponding to each of the sample decision words.

其中，预设词向量模型指的是用于将各词语表征为与其对应的向量编码的模型，示例性地，该预设词向量模型可以为word2vec模型。Among them, the preset word vector model refers to a model used to represent each word as a vector encoding corresponding to it. Illustratively, the preset word vector model may be a word2vec model.

具体地，在获取样本数据集之后，将各样本信息中的各样本决策词语输入至预设词向量模型中，对各样本决策词语进行向量表示，进而得到各样本信息中与各样本决策词语对应的决策词向量。Specifically, after obtaining the sample data set, each sample decision word in each sample information is input into the preset word vector model, and each sample decision word is vectorized, and then each sample information corresponding to each sample decision word is obtained. Decision word vector.

S207S207 ：获取与各决策词向量对应的样本权重，并根据各所述决策词向量对应的样本权重以及预设加权方法，确定所述预设分诊参考向量。: Obtain the sample weight corresponding to each decision word vector, and determine the preset triage reference vector according to the sample weight corresponding to each decision word vector and a preset weighting method.

其中，样本权重指的是各决策词向量在与其对应的样本信息中出现的频率以及各决策向量在样本数据集中出现的频次的综合权重。预设加权方法指的是加权平均方法。Among them, the sample weight refers to the comprehensive weight of the frequency of each decision word vector in the corresponding sample information and the frequency of each decision vector in the sample data set. The preset weighting method refers to the weighted average method.

具体地，在将各所述样本决策词语输入至预设词向量模型中，对各所述样本决策词语进行向量表示，得到与各所述样本决策词语对应的决策词向量之后，通过TF-IDF（Term Frequency–Inverse Document Frequency，词频-逆文本频率指数）技术，计算各决策词向量在与其对应的样本信息中出现的频率以及各决策向量在样本数据集中出现的频次的综合权重，进而得到与各决策词向量对应的样本权重；通过对各决策词向量对应的样本权重进行加权平均的方法，进而将各决策词向量统一表示为预设分诊参考向量。Specifically, after each of the sample decision words is input into the preset word vector model, each of the sample decision words is expressed in vector, and the decision word vector corresponding to each of the sample decision words is obtained, and then the TF-IDF (Term Frequency-Inverse Document Frequency, term frequency-inverse document frequency index) technology, calculates the frequency of each decision word vector in its corresponding sample information and the comprehensive weight of the frequency of each decision vector in the sample data set, and then obtains and The sample weight corresponding to each decision word vector; through the weighted average method of the sample weight corresponding to each decision word vector, each decision word vector is uniformly expressed as a preset triage reference vector.

S204S204 ：根据各所述实体向量与所述预设分诊参考向量之间的匹配值，得到与所述特征信息文本对应的所述解释结果。: Obtain the interpretation result corresponding to the characteristic information text according to the matching value between each of the entity vectors and the preset triage reference vector.

具体地，在获取预设分诊参考向量，将各所述实体向量与所述预设分诊参考向量输入至所述预设解释模型中的注意力模型中，确定各所述实体向量与所述预设分诊参考向量之间的匹配值之后，根据各所述实体向量与所述预设分诊参考向量之间的匹配值，得到与所述特征信息文本对应的所述解释结果。Specifically, after obtaining preset triage reference vectors, inputting each of the entity vectors and the preset triage reference vectors into the attention model in the preset interpretation model, and determining each of the entity vectors and the reference vectors After the matching value between the preset triage reference vectors is described, the interpretation result corresponding to the characteristic information text is obtained according to the matching value between each entity vector and the preset triage reference vector.

S30S30 ：比对所述特征信息文本的各词语对应的匹配值以及预设匹配阈值，确定所述特征信息文本中是否存在分诊解释词语；所述分诊解释词语是指所述特征信息文本中与大于或等于所述预设匹配阈值的匹配值对应的词语。: Compare the matching value corresponding to each word of the feature information text and the preset matching threshold to determine whether there are triage interpretation words in the feature information text; the triage interpretation words refer to the A word corresponding to a matching value greater than or equal to the preset matching threshold.

示例性地，预设匹配阈值可以为0.9，0.95等数值。Exemplarily, the preset matching threshold may be 0.9, 0.95, and other values.

具体地，在将所述特征信息文本输入至预设解释模型，对所述特征信息文本进行解释处理，得到与所述特征信息文本对应的解释结果；由于该解释结果中包含与特征信息文本中各词语对应的匹配值，进而比对各词语对应的匹配值以及预设匹配阈值，也即将各词语对应的匹配值与预设匹配阈值进行比较，以确定特征信息文本中是否存在分诊解释词语；Specifically, when the characteristic information text is input into a preset interpretation model, the characteristic information text is interpreted and processed to obtain an interpretation result corresponding to the characteristic information text; because the interpretation result includes the characteristic information text The matching value corresponding to each word is then compared with the matching value corresponding to each word and the preset matching threshold, that is, the matching value corresponding to each word is compared with the preset matching threshold to determine whether there are triage interpretation words in the feature information text ；

进一步地，在特征信息文本中，若存在大于或等于预设匹配阈值的匹配值对应的词语时，将该词语记录为分诊解释词语，该分诊解释词语用于对步骤S40中得到的分诊信息进行解释。Further, in the feature information text, if there is a word corresponding to a matching value greater than or equal to the preset matching threshold, the word is recorded as a triage explanation word, and the triage explanation word is used to compare the score obtained in step S40. Explain the diagnosis information.

S40S40 ：在所述特征信息文本中存在分诊解释词语时，将所述特征信息文本输入至预设分诊决策模型中，对所述特征信息文本进行分诊处理，得到与所述特征信息文本对应的分诊信息。: When there are triage interpretation words in the feature information text, input the feature information text into the preset triage decision model, and perform triage processing on the feature information text to obtain the text corresponding to the feature information Of triage information.

其中，预设分诊决策模型用于输出与特征信息文本对应的分诊信息，作为优选，该预设分诊决策模型可以采用非结构化的CNN（Convolutional Neural Networks，卷积神经网络）模型构建得到。分诊信息可以是药物推荐信息、也可以是科室推荐信息。Among them, the preset triage decision model is used to output the triage information corresponding to the feature information text. As a preference, the preset triage decision model can use unstructured CNN (Convolutional Neural Networks (convolutional neural network) model is constructed. The triage information can be drug recommendation information or department recommendation information.

具体地，在比对所述特征信息文本的各词语对应的匹配值以及预设匹配阈值，确定所述特征信息文本中是否存在分诊解释词语之后，在特征信息文本中存在分诊解释词语时，表征当前待分诊对象输入的特征信息文本中存在与医疗体系高度关联的词语，进而将特征信息文本输入至预设分诊决策模型中，对特征信息文本进行分类，得到与特征信息文本对应的分诊信息。由于特征信息文本中存在分诊解释词语，表明预设分诊决策模型可以识别其中与医疗体系相关的词语，进而提高了在对特征信息文本进行分诊处理后得到的分诊信息的准确率，并且该分诊解释词语可以对该分诊信息作出进一步解释。Specifically, after comparing the matching value corresponding to each word of the characteristic information text and the preset matching threshold, and determining whether there is a triage explanation word in the characteristic information text, when there is a triage explanation word in the characteristic information text , To characterize that there are words that are highly related to the medical system in the characteristic information text input by the current subject to be triaged, and then input the characteristic information text into the preset triage decision model, classify the characteristic information text, and obtain the corresponding characteristic information text Of triage information. The existence of triage interpretation words in the feature information text indicates that the preset triage decision model can identify the words related to the medical system, thereby improving the accuracy of triage information obtained after the feature information text is triaged. And the triage explanation word can further explain the triage information.

在一具体实施方式中，在步骤S30之后，也即在比对所述特征信息文本的各词语对应的匹配值以及预设匹配阈值，确定所述特征信息文本中是否存在分诊解释词语之后，还包括：In a specific embodiment, after step S30, that is, after comparing the matching value corresponding to each word of the characteristic information text and the preset matching threshold, it is determined whether there are triage interpretation words in the characteristic information text, Also includes:

在确定所述特征信息文本中并不存在分诊解释词语时，提示所述待分诊对象更新所述特征信息文本。When it is determined that there is no triage explanation word in the feature information text, the subject to be triaged is prompted to update the feature information text.

具体地，在特征信息文本中，若不存在大于或等于预设匹配阈值的匹配值对应的词语时（也即不存在分诊解释词语），表征当前待分诊对象输入的特征信息文本中没有与医疗体系关联的词语，也即在根据该特征信息文本进行确定分诊信息的准确率较低，因此不调用预设分诊决策模型对该特征信息文本进行分诊处理得到分诊信息，提示待分诊对象更新特征信息文本，以在更新后的特征信息文本中存在分诊解释词语时，将更新后的特征信息文版输入至预设分诊决策模型中。通过上述方法，可以在待分诊对象输入的特征信息文本中存在与医疗体系关联的词语时（如咳嗽、感冒等），才将特征信息文本输入至预设分诊决策模型给出具体的分诊信息，而不是在待分诊对象输入任何信息均调用预设分诊决策模型进行分诊处理，提高了预设分诊决策模型输出的分诊信息的准确率，并且每一分诊信息均有对应的分诊解释词语，可以让待分诊对象了解给出分诊信息的原因，提高了预设分诊决策模型的利用率以及效率。Specifically, in the feature information text, if there is no word corresponding to a matching value greater than or equal to the preset matching threshold (that is, there is no triage interpretation word), there is no feature information text input to characterize the current subject to be triaged. Words associated with the medical system, that is, the accuracy of determining triage information based on the feature information text is low, so the preset triage decision model is not called to triage the feature information text to obtain triage information, prompting The subject to be triaged updates the feature information text, so that when there are triage interpretation words in the updated feature information text, the updated feature information version is input into the preset triage decision model. Through the above method, only when there are words related to the medical system (such as cough, cold, etc.) in the feature information text input by the subject to be triaged, the feature information text can be input into the preset triage decision model to give specific scores. Instead of inputting any information in the subject to be triaged, the preset triage decision model is called for triage processing, which improves the accuracy of the triage information output by the preset triage decision model, and each triage information is There are corresponding triage explanation words, which can let the subjects to be triaged understand the reasons for giving triage information, and improve the utilization and efficiency of the preset triage decision-making model.

在一具体实施例中，如图4所示，步骤S40中，也即将所述特征信息文本输入至预设分诊决策模型中，对所述特征信息文本进行分诊处理，得到与所述特征信息文本对应的分诊信息，包括如下步骤：In a specific embodiment, as shown in FIG. 4, in step S40, the feature information text is input into the preset triage decision model, and the feature information text is triaged to obtain the feature information text. The triage information corresponding to the message text includes the following steps:

S401S401 ：获取所述特征信息文本中的结构化信息，通过所述预设分诊决策模型对所述结构化信息进行特征提取，得到与所述结构化信息对应的结构特征向量。: Acquire structured information in the feature information text, perform feature extraction on the structured information through the preset triage decision model, and obtain a structure feature vector corresponding to the structured information.

示例性地，结构化信息可以包括待分诊对象的基本信息（如年龄、性别等）或者统计学信息（如是否抽烟，是否饮酒等）。Exemplarily, the structured information may include basic information (such as age, gender, etc.) or statistical information (such as whether to smoke, whether to drink alcohol, etc.) of the subject to be triaged.

具体地，在所述特征信息文本中存在分诊解释词语时，获取特征信息文本中的结构化信息，通过预设分诊决策模型对结构化信息进行特征提取，进一步地可以通过预设分诊决策模型中的编码器对结构化信息进行特征向量抽取，得到与结构化信息对应的结构特征向量。Specifically, when there are triage interpretation words in the feature information text, the structured information in the feature information text is obtained, and the structured information is feature extracted through the preset triage decision model, and furthermore, the preset triage The encoder in the decision model extracts the feature vector of the structured information, and obtains the structure feature vector corresponding to the structured information.

S402S402 ：获取所述特征信息文本中的非结构化信息，确定所述特征信息文本中的非结构化信息的文本长度。: Obtain the unstructured information in the characteristic information text, and determine the text length of the unstructured information in the characteristic information text.

示例性地，非结构化信息可以为待分诊对象输入的病情信息（如发现咳嗽一周，并存在发烧现象等）。文本长度指的是非结构化信息中字符串总长度。Exemplarily, the unstructured information may be medical condition information input by the subject to be triaged (for example, a cough is found for a week, and there is a fever, etc.). The text length refers to the total length of the string in the unstructured information.

S403S403 ：根据所述文本长度以及预设历史文本长度，确定所述预设分诊决策模型中的卷积核类别。: Determine the convolution kernel category in the preset triage decision model according to the length of the text and the length of the preset historical text.

其中，预设历史文本可以为历史医疗数据（如住院信息文本）对应的文本，作为优选，预设历史文本长度选取各预设历史文本长度的平均值。Among them, the preset historical text may be a text corresponding to historical medical data (such as hospitalization information text). Preferably, the preset historical text length selects the average value of each preset historical text length.

具体地，在获取所述特征信息文本中的非结构化信息，确定所述特征信息文本中的非结构化信息的文本长度之后，根据文本长度以及预设历史文本长度，确定预设分诊决策模型中对非结构化信息进行特征识别的卷积核类别。Specifically, after obtaining the unstructured information in the characteristic information text and determining the text length of the unstructured information in the characteristic information text, determine the preset triage decision based on the text length and the preset historical text length The convolution kernel category for feature recognition of unstructured information in the model.

在一具体实施方式中，步骤S403包括如下步骤：In a specific implementation, step S403 includes the following steps:

在所述文本长度小于或等于所述预设历史文本长度时，将所述卷积核类别确定为小尺寸卷积核。When the text length is less than or equal to the preset historical text length, the convolution kernel category is determined as a small-size convolution kernel.

在所述文本长度大于所述预设历史文本长度时，将所述卷积核类别确定为大尺寸卷积核。When the text length is greater than the preset historical text length, the convolution kernel category is determined as a large-size convolution kernel.

其中，卷积核可以选取大小为2、3、4、5或者7等。假设本实施例中的预设分诊决策模型中包含四类大小的卷积核，具体为2、3、4和5。Among them, the size of the convolution kernel can be 2, 3, 4, 5, or 7, etc. It is assumed that the preset triage decision model in this embodiment includes four types of convolution kernels, specifically 2, 3, 4, and 5.

进一步地，在获取所述特征信息文本中的非结构化信息，确定所述特征信息文本中的非结构化信息的文本长度之后，将文本长度与预设历史文本长度进行比对，在文本长度小于或等于预设历史文本长度时，将卷积核类别确定为小尺寸卷积核，也即上述卷积核中大小为2、3、4的卷积核。由于文本长度较短（相较于预设历史文本长度）时，若采用尺寸太大的卷积核进行特征识别，可能无法抽取非结构化信息中的有效信息，并且可能会因为边界处的padding（填充）机制引入其它无效信息，降低了特征识别的准确率。Further, after obtaining the unstructured information in the characteristic information text and determining the text length of the unstructured information in the characteristic information text, the text length is compared with the preset historical text length, and the text length When the length is less than or equal to the preset historical text length, the convolution kernel category is determined as the small-size convolution kernel, that is, the convolution kernels with sizes 2, 3, and 4 in the above convolution kernels. Due to the short text length (compared to the preset historical text length), if a convolution kernel with a too large size is used for feature recognition, it may not be possible to extract valid information from unstructured information, and it may be due to padding at the boundary. The (filling) mechanism introduces other invalid information, which reduces the accuracy of feature recognition.

进一步地，在文本长度大于预设历史文本长度时，将卷积核类别确定为大尺寸卷积核，也即上述卷积核中大小为3、4、5的卷积核。由于文本长度较长（相较于预设历史文本长度）时，若采用尺寸较小的卷积核进行特征识别，可能会造成过于关注非结构化信息中的局部信息，而忽略了全局信息，导致特征识别过程中关注点发生错位，导致特征识别的准确率较低。Further, when the text length is greater than the preset historical text length, the convolution kernel type is determined to be a large-size convolution kernel, that is, the convolution kernels of sizes 3, 4, and 5 in the aforementioned convolution kernel. Due to the long text length (compared to the preset historical text length), if a smaller size convolution kernel is used for feature recognition, it may cause too much attention to the local information in the unstructured information, while ignoring the global information. As a result, the focus points are misplaced in the feature recognition process, resulting in low accuracy of feature recognition.

S404S404 ：通过所述卷积核类别中的卷积核对所述非结构化信息进行卷积池化操作，得到非结构特征向量。: Perform a convolution pooling operation on the unstructured information through the convolution kernel in the convolution kernel category to obtain an unstructured feature vector.

具体地，在根据所述文本长度以及预设历史文本长度，确定所述预设分诊决策模型中的卷积核类别之后，通过卷积核类别中的卷积核对非结构化信息进行卷积池化操作，以抽取非结构化信息中的特征，得到非结构特征向量。Specifically, after determining the convolution kernel category in the preset triage decision model according to the text length and the preset historical text length, unstructured information is convolved through the convolution kernel in the convolution kernel category Pooling operation to extract features from unstructured information to obtain unstructured feature vectors.

S405S405 ：对所述结构特征向量与所述非结构特征向量进行拼接，得到融合向量，并将所述融合向量输入至全连接层后，得到所述分诊信息。: Splicing the structural feature vector and the non-structural feature vector to obtain a fusion vector, and input the fusion vector to a fully connected layer to obtain the triage information.

其中，拼接的方式可以采用横向拼接也可以采用纵向拼接，作为优选，本实施例采用纵向拼接方式。Among them, the splicing method may be horizontal splicing or longitudinal splicing. As a preference, this embodiment adopts the longitudinal splicing method.

具体地，在通过所述卷积核类别中的卷积核对所述非结构化信息进行卷积池化操作，得到非结构特征向量之后，对结构特征向量与非结构特征向量进行纵向拼接，得到融合向量；将该融合向量输入至全连接层和softmax层进行分类，得到分诊信息。通过在预设分诊决策模型的隐层级别融合了结构化信息和非结构信息的特征向量，提升了预设分诊决策模型的分类精度，提高了分诊信息的准确率。Specifically, after performing a convolution pooling operation on the unstructured information through the convolution kernel in the convolution kernel category to obtain the unstructured feature vector, the structure feature vector and the unstructured feature vector are vertically spliced to obtain Fusion vector; input the fusion vector to the fully connected layer and the softmax layer for classification to obtain triage information. By fusing the feature vectors of structured information and unstructured information at the hidden level of the preset triage decision model, the classification accuracy of the preset triage decision model is improved, and the accuracy of triage information is improved.

S50S50 ：将所述分诊信息以及所述分诊解释词语关联推送至所述待分诊对象。: Push the triage information and the triage explanation word association to the object to be triaged.

具体地，在将所述特征信息文本输入至预设分诊决策模型中，对所述特征信息文本进行分诊处理，得到与所述特征信息文本对应的分诊信息之后，将分诊信息以及分诊解释词语关联推送至待分诊对象。Specifically, after inputting the feature information text into the preset triage decision model, performing triage processing on the feature information text, and after obtaining the triage information corresponding to the feature information text, the triage information and The triage interpretation words are associated and pushed to the subject to be triaged.

在本实施例中，通过引入解释模型，在特征信息文本中存在分诊解释词语时，才将特征信息文本输入至预设分诊决策模型，提高了预设分诊决策模型的利用率以及效率；并且通过将得到的分诊信息与分诊解释词语关联推送至待分诊对象，令待分诊对象了解给出该分诊信息的依据，在保证分诊信息的准确率的同时，提升了用户体验。In this embodiment, by introducing the interpretation model, the feature information text is input to the preset triage decision model only when there are triage interpretation words in the feature information text, which improves the utilization and efficiency of the preset triage decision model ; And by pushing the obtained triage information and triage explanation words to the object to be triaged, the object to be triaged can understand the basis for giving the triage information, while ensuring the accuracy of the triage information, it improves user experience.

在另一具体实施例中，为了保证上述实施例中的特征信息文本、预设解释模型以及预设分诊决策模型的私密以及安全性，可以将特征信息文本、预设解释模型以及预设分诊决策模型存储在区块链中。其中，区块链(Blockchain)，是由区块(Block)形成的加密的、链式的交易的存储结构。In another specific embodiment, in order to ensure the privacy and security of the feature information text, the preset interpretation model, and the preset triage decision model in the foregoing embodiment, the feature information text, the preset interpretation model, and the preset score may be The diagnosis decision model is stored in the blockchain. Among them, the Blockchain is an encrypted and chained transaction storage structure formed by blocks.

例如，每个区块的头部既可以包括区块中所有交易的哈希值，同时也包含前一个区块中所有交易的哈希值，从而基于哈希值实现区块中交易的防篡改和防伪造；新产生的交易被填充到区块并经过区块链网络中节点的共识后，会被追加到区块链的尾部从而形成链式的增长。For example, the header of each block can not only include the hash value of all transactions in the block, but also the hash value of all transactions in the previous block, so as to achieve tamper-proof transactions in the block based on the hash value And anti-counterfeiting; newly generated transactions are filled in the block and after the consensus of the nodes in the block chain network, they will be appended to the end of the block chain to form chain growth.

应理解，上述实施例中各步骤的序号的大小并不意味着执行顺序的先后，各过程的执行顺序应以其功能和内在逻辑确定，而不应对本申请实施例的实施过程构成任何限定。It should be understood that the size of the sequence number of each step in the foregoing embodiment does not mean the order of execution. The execution sequence of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present application.

在一实施例中，提供一种基于解释模型的分诊信息推荐装置，该基于解释模型的分诊信息推荐装置与上述实施例中基于解释模型的分诊信息推荐方法一一对应。如图5所示，该基于解释模型的分诊信息推荐装置包括文本获取模块10、解释处理模块20、匹配比对模块30、文本分类模块40以及分诊信息推送模块50。各功能模块详细说明如下：In one embodiment, an interpretation model-based triage information recommendation device is provided. The interpretation model-based triage information recommendation device corresponds to the interpretation model-based triage information recommendation method in the foregoing embodiment. As shown in FIG. 5, the triage information recommendation device based on the interpretation model includes a text acquisition module 10, an interpretation processing module 20, a matching and comparison module 30, a text classification module 40, and a triage information pushing module 50. The detailed description of each functional module is as follows:

文本获取模块10，用于获取待分诊对象录入的特征信息文本；The text acquisition module 10 is used to acquire the characteristic information text entered by the subject to be triaged;

解释处理模块20，用于将所述特征信息文本输入至预设解释模型，对所述特征信息文本进行解释处理，得到与所述特征信息文本对应的解释结果；所述解释结果中包含预设分诊参考向量与所述特征信息文本中各词语对应的匹配值；The interpretation processing module 20 is configured to input the characteristic information text into a preset interpretation model, perform interpretation processing on the characteristic information text, and obtain an interpretation result corresponding to the characteristic information text; the interpretation result includes the preset The matching value between the triage reference vector and each word in the characteristic information text;

匹配比对模块30，用于比对所述特征信息文本的各词语对应的匹配值以及预设匹配阈值，确定所述特征信息文本中是否存在分诊解释词语；所述分诊解释词语是指所述特征信息文本中与大于或等于所述预设匹配阈值的匹配值对应的词语。The matching and comparison module 30 is used to compare the matching value corresponding to each word of the characteristic information text and a preset matching threshold to determine whether there are triage interpretation words in the characteristic information text; the triage interpretation words refer to Words in the characteristic information text that correspond to a matching value greater than or equal to the preset matching threshold.

文本分类模块40，用于在所述特征信息文本中存在分诊解释词语时，将所述特征信息文本输入至预设分诊决策模型中，对所述特征信息文本进行分诊处理，得到与所述特征信息文本对应的分诊信息。The text classification module 40 is used to input the feature information text into a preset triage decision model when there are triage interpretation words in the feature information text, and perform triage processing on the feature information text to obtain the The triage information corresponding to the characteristic information text.

分诊信息推送模块50，用于将所述分诊信息以及所述分诊解释词语关联推送至所述待分诊对象。The triage information push module 50 is configured to push the triage information and the triage explanation word association to the object to be triaged.

优选地，如图6所示，解释处理模块20包括如下单元：Preferably, as shown in FIG. 6, the interpretation processing module 20 includes the following units:

分词处理单元201，用于根据预设词语实体库，对所述特征信息文本进行分词处理，得到与所述特征信息文本对应的各字词实体；The word segmentation processing unit 201 is configured to perform word segmentation processing on the characteristic information text according to a preset word entity database to obtain each word entity corresponding to the characteristic information text;

向量编码单元202，用于将各字词实体输入至所述预设解释模型的循环神经网络模型中，得到与各所述字词实体对应的实体向量；The vector encoding unit 202 is configured to input each word entity into the recurrent neural network model of the preset interpretation model to obtain an entity vector corresponding to each word entity;

匹配值Match value 确定单元203，用于获取预设分诊参考向量，将各所述实体向量与所述预设分诊参考向量输入至所述预设解释模型中的注意力模型中，确定各所述实体向量与所述预设分诊参考向量之间的匹配值；The determining unit 203 is configured to obtain a preset triage reference vector, input each entity vector and the preset triage reference vector into the attention model in the preset interpretation model, and determine each entity vector Matching value with the preset triage reference vector;

解释结果确定单元204，用于根据各所述实体向量与所述预设分诊参考向量之间的匹配值，得到所述解释结果。The interpretation result determination unit 204 is configured to obtain the interpretation result according to the matching value between each entity vector and the preset triage reference vector.

优选地，基于解释模型的分诊信息推荐装置还包括如下模块：Preferably, the triage information recommendation device based on the interpretation model further includes the following modules:

样本数据获取模块，用于获取样本数据集，所述样本数据集中包含至少一个样本信息；所述样本信息中包含至少一个样本决策词语；A sample data acquisition module for acquiring a sample data set, the sample data set contains at least one sample information; the sample information contains at least one sample decision word;

向量表征模块，用于将各所述样本决策词语输入至预设词向量模型中，对各所述样本决策词语进行向量表示，得到与各所述样本决策词语对应的决策词向量；The vector characterization module is used to input each of the sample decision words into a preset word vector model, and perform vector representation on each of the sample decision words to obtain a decision word vector corresponding to each of the sample decision words;

样本权重获取模块，用于获取与各决策词向量对应的样本权重，并根据各所述决策词向量对应的样本权重以及预设加权方法，确定所述预设分诊参考向量。The sample weight obtaining module is used to obtain the sample weight corresponding to each decision word vector, and determine the preset triage reference vector according to the sample weight corresponding to each decision word vector and a preset weighting method.

优选地，如图7所示，文本分类模块40包括如下单元：Preferably, as shown in FIG. 7, the text classification module 40 includes the following units:

结构化信息获取单元401，用于获取所述特征信息文本中的结构化信息，通过所述预设分诊决策模型对所述结构化信息进行特征提取，得到与所述结构化信息对应的结构特征向量；The structured information acquisition unit 401 is configured to acquire structured information in the feature information text, and perform feature extraction on the structured information through the preset triage decision model to obtain a structure corresponding to the structured information Feature vector;

非结构化信息获取单元402，用于获取所述特征信息文本中的非结构化信息，确定所述特征信息文本中的非结构化信息的文本长度；The unstructured information acquiring unit 402 is configured to acquire unstructured information in the characteristic information text, and determine the text length of the unstructured information in the characteristic information text;

卷积核选取单元403，用于根据所述文本长度以及预设历史文本长度，确定所述预设分诊决策模型中的卷积核类别；The convolution kernel selection unit 403 is configured to determine the convolution kernel category in the preset triage decision model according to the text length and the preset historical text length;

卷积池化单元Convolutional Pooling Unit 404404 ，用于通过所述卷积核类别中的卷积核对所述非结构化信息进行卷积池化操作，得到非结构特征向量；, For performing a convolution pooling operation on the unstructured information through a convolution kernel in the convolution kernel category to obtain an unstructured feature vector;

向量拼接单元405，用于对所述结构特征向量与所述非结构特征向量进行拼接，得到融合向量，并将所述融合向量输入至全连接层后，得到所述分诊信息。The vector splicing unit 405 is configured to splice the structural feature vector and the non-structural feature vector to obtain a fusion vector, and input the fusion vector to the fully connected layer to obtain the triage information.

优选地，卷积核选取单元403包括如下子单元：Preferably, the convolution kernel selection unit 403 includes the following subunits:

第一卷积核选取子单元，用于在所述文本长度小于或等于所述预设历史文本长度时，将所述卷积核类别确定为小尺寸卷积核；The first convolution kernel selection subunit is configured to determine the convolution kernel category as a small-size convolution kernel when the text length is less than or equal to the preset historical text length;

第二卷积核选取子单元，用于在所述文本长度大于所述预设历史文本长度时，将所述卷积核类别确定为大尺寸卷积核。The second convolution kernel selection subunit is configured to determine the convolution kernel category as a large-size convolution kernel when the text length is greater than the preset historical text length.

优选地，基于解释模型的分诊信息推荐装置还包括：Preferably, the triage information recommendation device based on the interpretation model further includes:

信息提示模块，用于在确定所述特征信息文本中并不存在分诊解释词语时，提示所述待分诊对象更新所述特征信息文本。The information prompting module is configured to prompt the subject to be triaged to update the characteristic information text when it is determined that there is no triage interpretation word in the characteristic information text.

关于基于解释模型的分诊信息推荐装置的具体限定可以参见上文中对于基于解释模型的分诊信息推荐方法的限定，在此不再赘述。上述基于解释模型的分诊信息推荐装置中的各个模块可全部或部分通过软件、硬件及其组合来实现。上述各模块可以硬件形式内嵌于或独立于计算机设备中的处理器中，也可以以软件形式存储于计算机设备中的存储器中，以便于处理器调用执行以上各个模块对应的操作。Regarding the specific limitation of the triage information recommendation device based on the interpretation model, please refer to the above limitation on the triage information recommendation method based on the interpretation model, which will not be repeated here. The various modules in the above-mentioned interpretation model-based triage information recommendation device can be implemented in whole or in part by software, hardware, and a combination thereof. The above-mentioned modules may be embedded in the form of hardware or independent of the processor in the computer equipment, or may be stored in the memory of the computer equipment in the form of software, so that the processor can call and execute the operations corresponding to the above-mentioned modules.

在一个实施例中，提供了一种计算机设备，该计算机设备可以是服务器，其内部结构图可以如图8所示。该计算机设备包括通过***总线连接的处理器、存储器、网络接口和数据库。其中，该计算机设备的处理器用于提供计算和控制能力。该计算机设备的存储器包括可读存储介质、内存储器。该可读存储介质存储有操作***、计算机可读指令和数据库。该内存储器为可读存储介质中的操作***和计算机可读指令的运行提供环境。该计算机设备的数据库用于存储上述实施例中基于解释模型的分诊信息推荐方法所使用到的数据。该计算机设备的网络接口用于与外部的终端通过网络连接通信。该计算机可读指令被处理器执行时以实现一种基于解释模型的分诊信息推荐方法。本实施例所提供的可读存储介质包括非易失性可读存储介质和易失性可读存储介质。In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. 8. The computer equipment includes a processor, a memory, a network interface, and a database connected through a system bus. Among them, the processor of the computer device is used to provide calculation and control capabilities. The memory of the computer device includes a readable storage medium and an internal memory. The readable storage medium stores an operating system, computer readable instructions, and a database. The internal memory provides an environment for the operation of the operating system and computer readable instructions in the readable storage medium. The database of the computer device is used to store the data used in the method for recommending triage information based on the interpretation model in the foregoing embodiment. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer-readable instructions are executed by the processor, a method for recommending triage information based on an interpretation model is realized. The readable storage medium provided in this embodiment includes a non-volatile readable storage medium and a volatile readable storage medium.

在一个实施例中，提供了一种计算机设备，包括存储器、处理器及存储在存储器上并可在处理器上运行的计算机可读指令，处理器执行计算机可读指令时实现如下步骤：In one embodiment, a computer device is provided, including a memory, a processor, and computer readable instructions stored in the memory and capable of running on the processor, and the processor implements the following steps when the processor executes the computer readable instructions:

获取待分诊对象录入的特征信息文本；Obtain the characteristic information text entered by the subject to be triaged;

将所述特征信息文本输入至预设解释模型，对所述特征信息文本进行解释处理，得到与所述特征信息文本对应的解释结果；所述解释结果中包含预设分诊参考向量与所述特征信息文本中各词语对应的匹配值；The feature information text is input into a preset interpretation model, and the feature information text is interpreted to obtain an interpretation result corresponding to the feature information text; the interpretation result includes the preset triage reference vector and the The matching value corresponding to each word in the feature information text;

比对所述特征信息文本的各词语对应的匹配值以及预设匹配阈值，确定所述特征信息文本中是否存在分诊解释词语；所述分诊解释词语是指所述特征信息文本中与大于或等于所述预设匹配阈值的匹配值对应的词语；Compare the matching value corresponding to each word of the characteristic information text and the preset matching threshold to determine whether there are triage interpretation words in the characteristic information text; Or a word corresponding to a matching value equal to the preset matching threshold;

在所述特征信息文本中存在分诊解释词语时，将所述特征信息文本输入至预设分诊决策模型中，对所述特征信息文本进行分诊处理，得到与所述特征信息文本对应的分诊信息；When there are triage interpretation words in the feature information text, input the feature information text into a preset triage decision model, and perform triage processing on the feature information text to obtain the corresponding feature information text Triage information;

将所述分诊信息以及所述分诊解释词语关联推送至所述待分诊对象。Push the triage information and the triage explanation word association to the object to be triaged.

在一个实施例中，提供了一个或多个存储有计算机可读指令的可读存储介质，本实施例所提供的可读存储介质包括非易失性可读存储介质和易失性可读存储介质；该可读存储介质上存储有计算机可读指令，该计算机可读指令被一个或多个处理器执行时，使得一个或多个处理器实现如下步骤：In one embodiment, one or more readable storage media storing computer readable instructions are provided. The readable storage media provided in this embodiment include non-volatile readable storage media and volatile readable storage. Medium; the readable storage medium stores computer readable instructions, and when the computer readable instructions are executed by one or more processors, the one or more processors implement the following steps:

获取待分诊对象录入的特征信息文本；Obtain the characteristic information text entered by the subject to be triaged;

将所述特征信息文本输入至预设解释模型，对所述特征信息文本进行解释处理，得到与所述特征信息文本对应的解释结果；所述解释结果中包含预设分诊参考向量与所述特征信息文本中各词语对应的匹配值；The feature information text is input into a preset interpretation model, and the feature information text is interpreted to obtain an interpretation result corresponding to the feature information text; the interpretation result includes the preset triage reference vector and the The matching value corresponding to each word in the feature information text;

比对所述特征信息文本的各词语对应的匹配值以及预设匹配阈值，确定所述特征信息文本中是否存在分诊解释词语；所述分诊解释词语是指所述特征信息文本中与大于或等于所述预设匹配阈值的匹配值对应的词语；Compare the matching value corresponding to each word of the characteristic information text and the preset matching threshold to determine whether there are triage interpretation words in the characteristic information text; Or a word corresponding to a matching value equal to the preset matching threshold;

在所述特征信息文本中存在分诊解释词语时，将所述特征信息文本输入至预设分诊决策模型中，对所述特征信息文本进行分诊处理，得到与所述特征信息文本对应的分诊信息；When there are triage interpretation words in the feature information text, input the feature information text into a preset triage decision model, and perform triage processing on the feature information text to obtain the corresponding feature information text Triage information;

将所述分诊信息以及所述分诊解释词语关联推送至所述待分诊对象。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，是可以通过计算机可读指令来指令相关的硬件来完成，所述的计算机可读指令可存储于一非易失性计算机可读取存储介质或者易失性计算机可读取存储介质中，该计算机可读指令在执行时，可包括如上述各方法的实施例的流程。其中，本申请所提供的各实施例中所使用的对存储器、存储、数据库或其它介质的任何引用，均可包括非易失性和/或易失性存储器。非易失性存储器可包括只读存储器（ROM）、可编程ROM（PROM）、电可编程ROM（EPROM）、电可擦除可编程ROM（EEPROM）或闪存。易失性存储器可包括随机存取存储器（RAM）或者外部高速缓冲存储器。作为说明而非局限，RAM以多种形式可得，诸如静态RAM（SRAM）、动态RAM（DRAM）、同步DRAM（SDRAM）、双数据率SDRAM（DDRSDRAM）、增强型SDRAM（ESDRAM）、同步链路（Synchlink） DRAM（SLDRAM）、存储器总线（Rambus）直接RAM（RDRAM）、直接存储器总线动态RAM（DRDRAM）、以及存储器总线动态RAM（RDRAM）等。Push the triage information and the triage explanation word association to the object to be triaged. A person of ordinary skill in the art can understand that all or part of the processes in the methods of the foregoing embodiments can be implemented by instructing relevant hardware through computer-readable instructions. The computer-readable instructions can be stored in a non-volatile computer. In a readable storage medium or a volatile computer readable storage medium, when the computer readable instruction is executed, it may include the processes of the above-mentioned method embodiments. Wherein, any reference to memory, storage, database, or other media used in the embodiments provided in this application may include non-volatile and/or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not a limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous chain Channel (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

所属领域的技术人员可以清楚地了解到，为了描述的方便和简洁，仅以上述各功能单元、模块的划分进行举例说明，实际应用中，可以根据需要而将上述功能分配由不同的功能单元、模块完成，即将所述装置的内部结构划分成不同的功能单元或模块，以完成以上描述的全部或者部分功能。Those skilled in the art can clearly understand that, for the convenience and conciseness of description, only the division of the above functional units and modules is used as an example. In practical applications, the above functions can be allocated to different functional units and modules as needed. Module completion, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above.

以上所述实施例仅用以说明本申请的技术方案，而非对其限制；尽管参照前述实施例对本申请进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本申请各实施例技术方案的精神和范围，均应包含在本申请的保护范围之内。 The above-mentioned embodiments are only used to illustrate the technical solutions of the present application, not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still implement the foregoing The technical solutions recorded in the examples are modified, or some of the technical features are equivalently replaced; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the application, and should be included in Within the scope of protection of this application .

Claims (20)

1. 一种基于解释模型的分诊信息推荐方法，其中，包括： A method for recommending triage information based on an interpretation model, which includes:

   获取待分诊对象录入的特征信息文本；Obtain the characteristic information text entered by the subject to be triaged;

   将所述特征信息文本输入至预设解释模型，对所述特征信息文本进行解释处理，得到与所述特征信息文本对应的解释结果；所述解释结果中包含预设分诊参考向量与所述特征信息文本中各词语对应的匹配值；The feature information text is input into a preset interpretation model, and the feature information text is interpreted to obtain an interpretation result corresponding to the feature information text; the interpretation result includes the preset triage reference vector and the The matching value corresponding to each word in the feature information text;

   比对所述特征信息文本的各词语对应的匹配值以及预设匹配阈值，确定所述特征信息文本中是否存在分诊解释词语；所述分诊解释词语是指所述特征信息文本中与大于或等于所述预设匹配阈值的匹配值对应的词语；Compare the matching value corresponding to each word of the characteristic information text and the preset matching threshold to determine whether there are triage interpretation words in the characteristic information text; Or a word corresponding to a matching value equal to the preset matching threshold;

   在所述特征信息文本中存在分诊解释词语时，将所述特征信息文本输入至预设分诊决策模型中，对所述特征信息文本进行分诊处理，得到与所述特征信息文本对应的分诊信息；When there are triage interpretation words in the feature information text, input the feature information text into a preset triage decision model, and perform triage processing on the feature information text to obtain the corresponding feature information text Triage information;

   将所述分诊信息以及所述分诊解释词语关联推送至所述待分诊对象。Push the triage information and the triage explanation word association to the object to be triaged.
2. 如权利要求1所述的基于解释模型的分诊信息推荐方法，其中，所述将所述特征信息文本输入至预设解释模型，对所述特征信息文本进行解释处理，得到与所述特征信息文本对应的解释结果，包括： The method for recommending triage information based on an interpretation model according to claim 1, wherein said inputting said characteristic information text into a preset interpretation model, and performing interpretation processing on said characteristic information text, to obtain information related to said characteristic information. Interpretation results corresponding to the text, including:

   根据预设词语实体库，对所述特征信息文本进行分词处理，得到与所述特征信息文本对应的各字词实体；Performing word segmentation processing on the characteristic information text according to a preset word entity database to obtain each word entity corresponding to the characteristic information text;

   将各字词实体输入至所述预设解释模型的循环神经网络模型中，得到与各所述字词实体对应的实体向量；Input each word entity into the recurrent neural network model of the preset interpretation model to obtain an entity vector corresponding to each word entity;

   获取预设分诊参考向量，将各所述实体向量与所述预设分诊参考向量输入至所述预设解释模型中的注意力模型中，确定各所述实体向量与所述预设分诊参考向量之间的匹配值；Obtain a preset triage reference vector, input each of the entity vectors and the preset triage reference vector into the attention model in the preset interpretation model, and determine each of the entity vectors and the preset score The matching value between the diagnosis reference vectors;

   根据各所述实体向量与所述预设分诊参考向量之间的匹配值，得到与所述特征信息文本对应的所述解释结果。According to the matching value between each entity vector and the preset triage reference vector, the interpretation result corresponding to the characteristic information text is obtained.
3. 如权利要求2所述的基于解释模型的分诊信息推荐方法，其中，所述获取预设分诊参考向量之前，还包括： The method for recommending triage information based on an interpretation model according to claim 2, wherein before said obtaining a preset triage reference vector, the method further comprises:

   获取样本数据集，所述样本数据集中包含至少一个样本信息；所述样本信息中包含至少一个样本决策词语；Acquiring a sample data set, where the sample data set includes at least one sample information; the sample information includes at least one sample decision word;

   将各所述样本决策词语输入至预设词向量模型中，对各所述样本决策词语进行向量表示，得到与各所述样本决策词语对应的决策词向量；Input each of the sample decision words into a preset word vector model, and perform a vector representation on each of the sample decision words to obtain a decision word vector corresponding to each of the sample decision words;

   获取与各决策词向量对应的样本权重，并根据各所述决策词向量对应的样本权重以及预设加权方法，确定所述预设分诊参考向量。The sample weight corresponding to each decision word vector is acquired, and the preset triage reference vector is determined according to the sample weight corresponding to each decision word vector and a preset weighting method.
4. 如权利要求1所述的基于解释模型的分诊信息推荐方法，其中，所述将所述特征信息文本输入至预设分诊决策模型中，对所述特征信息文本进行分诊处理，得到与所述特征信息文本对应的分诊信息，包括： The method for recommending triage information based on an interpretation model according to claim 1, wherein said inputting said characteristic information text into a preset triage decision model, and performing triage processing on said characteristic information text to obtain and The triage information corresponding to the characteristic information text includes:

   获取所述特征信息文本中的结构化信息，通过所述预设分诊决策模型对所述结构化信息进行特征提取，得到与所述结构化信息对应的结构特征向量；Acquiring the structured information in the feature information text, and performing feature extraction on the structured information through the preset triage decision model to obtain a structure feature vector corresponding to the structured information;

   获取所述特征信息文本中的非结构化信息，确定所述特征信息文本中的非结构化信息的文本长度；Acquiring the unstructured information in the characteristic information text, and determining the text length of the unstructured information in the characteristic information text;

   根据所述文本长度以及预设历史文本长度，确定所述预设分诊决策模型中的卷积核类别；Determine the convolution kernel category in the preset triage decision model according to the text length and the preset historical text length;

   通过所述卷积核类别中的卷积核对所述非结构化信息进行卷积池化操作，得到非结构特征向量；Performing a convolution pooling operation on the unstructured information by using a convolution kernel in the convolution kernel category to obtain an unstructured feature vector;

   对所述结构特征向量与所述非结构特征向量进行拼接，得到融合向量，并将所述融合向量输入至全连接层后，得到所述分诊信息。The structural feature vector and the non-structural feature vector are spliced to obtain a fusion vector, and after the fusion vector is input to a fully connected layer, the triage information is obtained.
5. 如权利要求4所述的基于解释模型的分诊信息推荐方法，其中，所述根据所述文本长度以及预设历史文本长度，确定所述预设分诊决策模型中的卷积核类别，包括： The method for recommending triage information based on an interpretation model according to claim 4, wherein the determining the convolution kernel category in the preset triage decision model according to the text length and a preset historical text length includes :

   在所述文本长度小于或等于所述预设历史文本长度时，将所述卷积核类别确定为小尺寸卷积核；When the text length is less than or equal to the preset historical text length, determining the convolution kernel category as a small-size convolution kernel;

   在所述文本长度大于所述预设历史文本长度时，将所述卷积核类别确定为大尺寸卷积核。When the text length is greater than the preset historical text length, the convolution kernel category is determined as a large-size convolution kernel.
6. 如权利要求1所述的基于解释模型的分诊信息推荐方法，其中，所述确定所述特征信息文本中是否存在分诊解释词语之后，还包括： The method for recommending triage information based on an interpretation model according to claim 1, wherein after determining whether there are triage interpretation words in the characteristic information text, the method further comprises:

   在确定所述特征信息文本中并不存在分诊解释词语时，提示所述待分诊对象更新所述特征信息文本。When it is determined that there is no triage explanation word in the feature information text, the subject to be triaged is prompted to update the feature information text.
7. 一种基于解释模型的分诊信息推荐装置，其中，包括： An interpretation model-based triage information recommendation device, which includes:

   文本获取模块，用于获取待分诊对象录入的特征信息文本；The text acquisition module is used to acquire the characteristic information text entered by the object to be triaged;

   解释处理模块，用于将所述特征信息文本输入至预设解释模型，对所述特征信息文本进行解释处理，得到与所述特征信息文本对应的解释结果；所述解释结果中包含预设分诊参考向量与所述特征信息文本中各词语对应的匹配值；The interpretation processing module is used to input the feature information text into a preset interpretation model, perform interpretation processing on the feature information text, and obtain an interpretation result corresponding to the feature information text; the interpretation result includes the preset score The matching value between the diagnosis reference vector and each word in the characteristic information text;

   匹配比对模块，用于比对所述特征信息文本的各词语对应的匹配值以及预设匹配阈值，确定所述特征信息文本中是否存在分诊解释词语；所述分诊解释词语是指所述特征信息文本中与大于或等于所述预设匹配阈值的匹配值对应的词语；The matching and comparison module is used to compare the matching value corresponding to each word of the characteristic information text and the preset matching threshold to determine whether there are triage interpretation words in the characteristic information text; the triage interpretation words refer to all Words in the feature information text that correspond to a matching value greater than or equal to the preset matching threshold;

   文本分类模块，用于在所述特征信息文本中存在分诊解释词语时，将所述特征信息文本输入至预设分诊决策模型中，对所述特征信息文本进行分诊处理，得到与所述特征信息文本对应的分诊信息；The text classification module is used to input the feature information text into the preset triage decision model when there are triage interpretation words in the feature information text, and perform triage processing on the feature information text to obtain the The triage information corresponding to the text of the characteristic information;

   分诊信息推送模块，用于将所述分诊信息以及所述分诊解释词语关联推送至所述待分诊对象。The triage information push module is configured to push the triage information and the triage explanation word association to the object to be triaged.
8. 如权利要求7所述的基于解释模型的分诊信息推荐装置，其中，所述解释处理模块包括： The device for recommending triage information based on an interpretation model according to claim 7, wherein the interpretation processing module comprises:

   分词处理单元，用于根据预设词语实体库，对所述特征信息文本进行分词处理，得到与所述特征信息文本对应的各字词实体；The word segmentation processing unit is configured to perform word segmentation processing on the characteristic information text according to a preset word entity database to obtain each word entity corresponding to the characteristic information text;

   向量编码单元，用于将各字词实体输入至所述预设解释模型的循环神经网络模型中，得到与各所述字词实体对应的实体向量；The vector encoding unit is used to input each word entity into the cyclic neural network model of the preset interpretation model to obtain an entity vector corresponding to each word entity;

   匹配值确定单元，用于获取预设分诊参考向量，将各所述实体向量与所述预设分诊参考向量输入至所述预设解释模型中的注意力模型中，确定各所述实体向量与所述预设分诊参考向量之间的匹配值；The matching value determination unit is used to obtain a preset triage reference vector, input each entity vector and the preset triage reference vector into the attention model in the preset interpretation model, and determine each entity Matching value between the vector and the preset triage reference vector;

   解释结果确定单元，用于根据各所述实体向量与所述预设分诊参考向量之间的匹配值，得到所述解释结果。The interpretation result determination unit is configured to obtain the interpretation result according to the matching value between each entity vector and the preset triage reference vector.
9. 一种计算机设备，包括存储器、处理器以及存储在所述存储器中并可在所述处理器上运行的计算机可读指令，其中，所述处理器执行所述计算机可读指令时实现如下步骤： A computer device includes a memory, a processor, and computer-readable instructions that are stored in the memory and can run on the processor, wherein the processor implements the following steps when the processor executes the computer-readable instructions:

   获取待分诊对象录入的特征信息文本；Obtain the characteristic information text entered by the subject to be triaged;

   将所述特征信息文本输入至预设解释模型，对所述特征信息文本进行解释处理，得到与所述特征信息文本对应的解释结果；所述解释结果中包含预设分诊参考向量与所述特征信息文本中各词语对应的匹配值；The feature information text is input into a preset interpretation model, and the feature information text is interpreted to obtain an interpretation result corresponding to the feature information text; the interpretation result includes the preset triage reference vector and the The matching value corresponding to each word in the feature information text;

   比对所述特征信息文本的各词语对应的匹配值以及预设匹配阈值，确定所述特征信息文本中是否存在分诊解释词语；所述分诊解释词语是指所述特征信息文本中与大于或等于所述预设匹配阈值的匹配值对应的词语；Compare the matching value corresponding to each word of the characteristic information text and the preset matching threshold to determine whether there are triage interpretation words in the characteristic information text; Or a word corresponding to a matching value equal to the preset matching threshold;

   在所述特征信息文本中存在分诊解释词语时，将所述特征信息文本输入至预设分诊决策模型中，对所述特征信息文本进行分诊处理，得到与所述特征信息文本对应的分诊信息；When there are triage interpretation words in the feature information text, input the feature information text into a preset triage decision model, and perform triage processing on the feature information text to obtain the corresponding feature information text Triage information;

   将所述分诊信息以及所述分诊解释词语关联推送至所述待分诊对象。Push the triage information and the triage explanation word association to the object to be triaged.
10. 如权利要求9所述的计算机设备，其中，所述将所述特征信息文本输入至预设解释模型，对所述特征信息文本进行解释处理，得到与所述特征信息文本对应的解释结果，包括：The computer device according to claim 9, wherein said inputting said characteristic information text into a preset interpretation model, performing interpretation processing on said characteristic information text, and obtaining an interpretation result corresponding to said characteristic information text comprises :

    根据预设词语实体库，对所述特征信息文本进行分词处理，得到与所述特征信息文本对应的各字词实体；Performing word segmentation processing on the characteristic information text according to a preset word entity database to obtain each word entity corresponding to the characteristic information text;

    将各字词实体输入至所述预设解释模型的循环神经网络模型中，得到与各所述字词实体对应的实体向量；Input each word entity into the recurrent neural network model of the preset interpretation model to obtain an entity vector corresponding to each word entity;

    获取预设分诊参考向量，将各所述实体向量与所述预设分诊参考向量输入至所述预设解释模型中的注意力模型中，确定各所述实体向量与所述预设分诊参考向量之间的匹配值；Obtain a preset triage reference vector, input each of the entity vectors and the preset triage reference vector into the attention model in the preset interpretation model, and determine each of the entity vectors and the preset score The matching value between the diagnosis reference vectors;

    根据各所述实体向量与所述预设分诊参考向量之间的匹配值，得到与所述特征信息文本对应的所述解释结果。According to the matching value between each entity vector and the preset triage reference vector, the interpretation result corresponding to the characteristic information text is obtained.
11. 如权利要求10所述的计算机设备，其中，所述获取预设分诊参考向量之前，所述处理器执行所述计算机可读指令时还实现如下步骤：10. The computer device according to claim 10, wherein, before said obtaining the preset triage reference vector, the processor further implements the following steps when executing the computer readable instruction:

    获取样本数据集，所述样本数据集中包含至少一个样本信息；所述样本信息中包含至少一个样本决策词语；Acquiring a sample data set, where the sample data set includes at least one sample information; the sample information includes at least one sample decision word;

    将各所述样本决策词语输入至预设词向量模型中，对各所述样本决策词语进行向量表示，得到与各所述样本决策词语对应的决策词向量；Input each of the sample decision words into a preset word vector model, and perform a vector representation on each of the sample decision words to obtain a decision word vector corresponding to each of the sample decision words;

    获取与各决策词向量对应的样本权重，并根据各所述决策词向量对应的样本权重以及预设加权方法，确定所述预设分诊参考向量。The sample weight corresponding to each decision word vector is acquired, and the preset triage reference vector is determined according to the sample weight corresponding to each decision word vector and a preset weighting method.
12. 如权利要求9所述的计算机设备，其中，所述将所述特征信息文本输入至预设分诊决策模型中，对所述特征信息文本进行分诊处理，得到与所述特征信息文本对应的分诊信息，包括：The computer device according to claim 9, wherein said inputting said characteristic information text into a preset triage decision model, and performing triage processing on said characteristic information text, to obtain the corresponding characteristic information text Triage information, including:

    获取所述特征信息文本中的结构化信息，通过所述预设分诊决策模型对所述结构化信息进行特征提取，得到与所述结构化信息对应的结构特征向量；Acquiring the structured information in the feature information text, and performing feature extraction on the structured information through the preset triage decision model to obtain a structure feature vector corresponding to the structured information;

    获取所述特征信息文本中的非结构化信息，确定所述特征信息文本中的非结构化信息的文本长度；Acquiring the unstructured information in the characteristic information text, and determining the text length of the unstructured information in the characteristic information text;

    根据所述文本长度以及预设历史文本长度，确定所述预设分诊决策模型中的卷积核类别；Determine the convolution kernel category in the preset triage decision model according to the text length and the preset historical text length;

    通过所述卷积核类别中的卷积核对所述非结构化信息进行卷积池化操作，得到非结构特征向量；Performing a convolution pooling operation on the unstructured information by using a convolution kernel in the convolution kernel category to obtain an unstructured feature vector;

    对所述结构特征向量与所述非结构特征向量进行拼接，得到融合向量，并将所述融合向量输入至全连接层后，得到所述分诊信息。The structural feature vector and the non-structural feature vector are spliced to obtain a fusion vector, and after the fusion vector is input to a fully connected layer, the triage information is obtained.
13. 如权利要求12所述的计算机设备，其中，所述根据所述文本长度以及预设历史文本长度，确定所述预设分诊决策模型中的卷积核类别，包括：The computer device of claim 12, wherein the determining the convolution kernel category in the preset triage decision model according to the text length and a preset historical text length comprises:

    在所述文本长度小于或等于所述预设历史文本长度时，将所述卷积核类别确定为小尺寸卷积核；When the text length is less than or equal to the preset historical text length, determining the convolution kernel category as a small-size convolution kernel;

    在所述文本长度大于所述预设历史文本长度时，将所述卷积核类别确定为大尺寸卷积核。When the text length is greater than the preset historical text length, the convolution kernel category is determined as a large-size convolution kernel.
14. 如权利要求9所述的计算机设备，其中，所述确定所述特征信息文本中是否存在分诊解释词语之后，所述处理器执行所述计算机可读指令时还实现如下步骤：9. The computer device according to claim 9, wherein, after said determining whether there are triage interpretation words in the characteristic information text, the processor further implements the following steps when executing the computer-readable instructions:

    在确定所述特征信息文本中并不存在分诊解释词语时，提示所述待分诊对象更新所述特征信息文本。When it is determined that there is no triage explanation word in the feature information text, the subject to be triaged is prompted to update the feature information text.
15. 一个或多个存储有计算机可读指令的可读存储介质，其中，所述计算机可读指令被一个或多个处理器执行时，使得所述一个或多个处理器执行如下步骤： One or more readable storage media storing computer readable instructions, where when the computer readable instructions are executed by one or more processors, the one or more processors execute the following steps:

    获取待分诊对象录入的特征信息文本；Obtain the characteristic information text entered by the subject to be triaged;

    将所述特征信息文本输入至预设解释模型，对所述特征信息文本进行解释处理，得到与所述特征信息文本对应的解释结果；所述解释结果中包含预设分诊参考向量与所述特征信息文本中各词语对应的匹配值；The feature information text is input into a preset interpretation model, and the feature information text is interpreted to obtain an interpretation result corresponding to the feature information text; the interpretation result includes the preset triage reference vector and the The matching value corresponding to each word in the feature information text;

    比对所述特征信息文本的各词语对应的匹配值以及预设匹配阈值，确定所述特征信息文本中是否存在分诊解释词语；所述分诊解释词语是指所述特征信息文本中与大于或等于所述预设匹配阈值的匹配值对应的词语；Compare the matching value corresponding to each word of the characteristic information text and the preset matching threshold to determine whether there are triage interpretation words in the characteristic information text; Or a word corresponding to a matching value equal to the preset matching threshold;

    在所述特征信息文本中存在分诊解释词语时，将所述特征信息文本输入至预设分诊决策模型中，对所述特征信息文本进行分诊处理，得到与所述特征信息文本对应的分诊信息；When there are triage interpretation words in the feature information text, input the feature information text into a preset triage decision model, and perform triage processing on the feature information text to obtain the corresponding feature information text Triage information;

    将所述分诊信息以及所述分诊解释词语关联推送至所述待分诊对象。Push the triage information and the triage explanation word association to the object to be triaged.
16. 如权利要求15所述的可读存储介质，其中，所述将所述特征信息文本输入至预设解释模型，对所述特征信息文本进行解释处理，得到与所述特征信息文本对应的解释结果，包括：The readable storage medium according to claim 15, wherein said inputting said characteristic information text into a preset interpretation model, performing interpretation processing on said characteristic information text, and obtaining an interpretation result corresponding to said characteristic information text ,include:

    根据预设词语实体库，对所述特征信息文本进行分词处理，得到与所述特征信息文本对应的各字词实体；Performing word segmentation processing on the characteristic information text according to a preset word entity database to obtain each word entity corresponding to the characteristic information text;

    将各字词实体输入至所述预设解释模型的循环神经网络模型中，得到与各所述字词实体对应的实体向量；Input each word entity into the recurrent neural network model of the preset interpretation model to obtain an entity vector corresponding to each word entity;

    获取预设分诊参考向量，将各所述实体向量与所述预设分诊参考向量输入至所述预设解释模型中的注意力模型中，确定各所述实体向量与所述预设分诊参考向量之间的匹配值；Obtain a preset triage reference vector, input each of the entity vectors and the preset triage reference vector into the attention model in the preset interpretation model, and determine each of the entity vectors and the preset score The matching value between the diagnosis reference vectors;

    根据各所述实体向量与所述预设分诊参考向量之间的匹配值，得到与所述特征信息文本对应的所述解释结果。According to the matching value between each entity vector and the preset triage reference vector, the interpretation result corresponding to the characteristic information text is obtained.
17. 如权利要求16所述的可读存储介质，其中，所述获取预设分诊参考向量之前，所述计算机可读指令被一个或多个处理器执行时，使得所述一个或多个处理器还执行如下步骤：The readable storage medium according to claim 16, wherein, before the acquisition of the preset triage reference vector, when the computer-readable instruction is executed by one or more processors, the one or more processors are Also perform the following steps:

    获取样本数据集，所述样本数据集中包含至少一个样本信息；所述样本信息中包含至少一个样本决策词语；Acquiring a sample data set, where the sample data set includes at least one sample information; the sample information includes at least one sample decision word;

    将各所述样本决策词语输入至预设词向量模型中，对各所述样本决策词语进行向量表示，得到与各所述样本决策词语对应的决策词向量；Input each of the sample decision words into a preset word vector model, and perform a vector representation on each of the sample decision words to obtain a decision word vector corresponding to each of the sample decision words;

    获取与各决策词向量对应的样本权重，并根据各所述决策词向量对应的样本权重以及预设加权方法，确定所述预设分诊参考向量。The sample weight corresponding to each decision word vector is acquired, and the preset triage reference vector is determined according to the sample weight corresponding to each decision word vector and a preset weighting method.
18. 如权利要求15所述的可读存储介质，其中，所述将所述特征信息文本输入至预设分诊决策模型中，对所述特征信息文本进行分诊处理，得到与所述特征信息文本对应的分诊信息，包括：The readable storage medium according to claim 15, wherein said inputting said characteristic information text into a preset triage decision model, performing triage processing on said characteristic information text, and obtaining the same as said characteristic information text The corresponding triage information includes:

    获取所述特征信息文本中的结构化信息，通过所述预设分诊决策模型对所述结构化信息进行特征提取，得到与所述结构化信息对应的结构特征向量；Acquiring the structured information in the feature information text, and performing feature extraction on the structured information through the preset triage decision model to obtain a structure feature vector corresponding to the structured information;

    获取所述特征信息文本中的非结构化信息，确定所述特征信息文本中的非结构化信息的文本长度；Acquiring the unstructured information in the characteristic information text, and determining the text length of the unstructured information in the characteristic information text;

    根据所述文本长度以及预设历史文本长度，确定所述预设分诊决策模型中的卷积核类别；Determine the convolution kernel category in the preset triage decision model according to the text length and the preset historical text length;

    通过所述卷积核类别中的卷积核对所述非结构化信息进行卷积池化操作，得到非结构特征向量；Performing a convolution pooling operation on the unstructured information by using a convolution kernel in the convolution kernel category to obtain an unstructured feature vector;

    对所述结构特征向量与所述非结构特征向量进行拼接，得到融合向量，并将所述融合向量输入至全连接层后，得到所述分诊信息。The structural feature vector and the non-structural feature vector are spliced to obtain a fusion vector, and after the fusion vector is input to a fully connected layer, the triage information is obtained.
19. 如权利要求18所述的可读存储介质，其中，所述根据所述文本长度以及预设历史文本长度，确定所述预设分诊决策模型中的卷积核类别，包括：The readable storage medium of claim 18, wherein the determining the convolution kernel category in the preset triage decision model according to the text length and the preset historical text length comprises:

    在所述文本长度小于或等于所述预设历史文本长度时，将所述卷积核类别确定为小尺寸卷积核；When the text length is less than or equal to the preset historical text length, determining the convolution kernel category as a small-size convolution kernel;

    在所述文本长度大于所述预设历史文本长度时，将所述卷积核类别确定为大尺寸卷积核。When the text length is greater than the preset historical text length, the convolution kernel category is determined as a large-size convolution kernel.
20. 如权利要求15所述的可读存储介质，其中，所述确定所述特征信息文本中是否存在分诊解释词语之后，所述计算机可读指令被一个或多个处理器执行时，使得所述一个或多个处理器还执行如下步骤：The readable storage medium according to claim 15, wherein after said determining whether there are triage interpretation words in the characteristic information text, when the computer-readable instructions are executed by one or more processors, the One or more processors also perform the following steps:

    在确定所述特征信息文本中并不存在分诊解释词语时，提示所述待分诊对象更新所述特征信息文本。When it is determined that there is no triage explanation word in the feature information text, the subject to be triaged is prompted to update the feature information text.