CN111339252A

CN111339252A - Searching method, searching device and storage medium

Info

Publication number: CN111339252A
Application number: CN202010115111.2A
Authority: CN
Inventors: 文瑞; 陈曦; 高文龙; 孙继超; 赵博; 刘羽
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2020-02-25
Filing date: 2020-02-25
Publication date: 2020-06-26
Anticipated expiration: 2040-02-25
Also published as: CN111339252B

Abstract

The embodiment of the application provides a searching method, a searching device and a storage medium, wherein the method comprises the following steps: acquiring search information; obtaining at least one symptom feature based on the search information; setting weights for all symptom characteristics according to the semantic association degree of the symptom characteristics and symptoms; acquiring target similarity between each symptom characteristic with the set weight and at least one preset symptom entity; and determining a target symptom entity from the at least one preset symptom entity according to the target similarity and outputting the target symptom entity. By the scheme, the accuracy of symptom entities can be identified, and similar symptoms and non-similar symptoms can be effectively distinguished.

Description

Searching method, searching device and storage medium

Technical Field

The embodiment of the application relates to the technical field of artificial intelligence, in particular to a searching method, a searching device and a storage medium.

Background

In the medical field, artificial intelligence is commonly used to provide self-service medical services to users. For example, in a medical intelligent assistant application implemented based on artificial intelligence, the medical intelligent assistant application performs services such as subsequent disease judgment, department guidance, and doctor finding by registration according to input symptoms of a user. However, the user's input is often very different for the same symptom. For example, again the standard symptom "abdominal pain", user a's input may be: "the belly is a bit painful", the input of user b might be: "the belly is light and slightly painful". Since subsequent procedures such as disease judgment, department guidance, registration for finding doctors, etc. need to be performed according to the symptoms input by the user, the medical intelligent assistant application needs to identify similar symptoms with higher accuracy and link the input symptoms to a standard symptom library.

In the research and practice process of the prior art, the inventor of the embodiment of the application finds that the model in the medical intelligent assistant application is obtained by training a supervised machine learning or deep learning classification model based on training data of artificial labeling, and although the generalization capability of the model obtained by final training is strong, the model needs to depend on a large amount of accurate artificial labeling data, and the artificial labeling cost is high.

Disclosure of Invention

The embodiment of the application provides a searching method, a searching device and a searching medium, which can improve the accuracy of identifying symptom entities, effectively distinguish similar symptoms from non-similar symptoms, and do not depend on a large number of manual labels, thereby reducing the workload.

In a first aspect, an embodiment of the present application provides a search method, where the method includes:

acquiring search information;

obtaining at least one symptom feature based on the search information;

setting weights for all symptom characteristics according to the semantic association degree of the symptom characteristics and symptoms;

acquiring target similarity between each symptom characteristic with the set weight and at least one preset symptom entity;

and determining a target symptom entity from the at least one preset symptom entity according to the target similarity and outputting the target symptom entity.

In one possible design, the setting a weight for each symptom feature according to the semantic association degree between the symptom feature and the symptom includes:

determining symptom characteristics having semantic relation with symptoms according to the semantics of the symptom characteristics;

determining semantic association degree of each symptom characteristic and symptom according to the semantic relation;

and respectively setting the weight which is in positive relation with the semantic relevance for each symptom characteristic.

In one possible design, the preset symptom entity corresponds to at least one symptom description information; the obtaining of the target similarity between each symptom feature and a preset symptom entity includes:

respectively calculating the weighted similarity between each symptom characteristic and each symptom description information;

and taking the average value of the weighted similarity between each symptom characteristic and each symptom description information, and taking the average value as the target similarity.

In one possible design, the deriving at least one symptom feature based on the search information includes:

outputting symptom recommendation information obtained according to the search information, wherein the symptom recommendation information corresponds to a preset symptom entity;

acquiring operation behavior data of the user on the symptom recommendation information;

and obtaining the at least one symptom characteristic according to the operation behavior data and the search information.

In one possible design, the determining a target symptom entity from the at least one preset symptom entity according to the target similarity includes:

obtaining an operation behavior map according to the operation behavior data;

determining candidate symptom entities according to the operation behavior map and a preset threshold;

obtaining the similarity between the candidate symptom entity and at least one symptom description information;

and determining the target symptom entity according to the similarity.

In one possible design, the method further includes:

acquiring training data, wherein the training data comprises a plurality of medical corpora;

determining a plurality of medical symptoms according to the plurality of medical corpora;

respectively converting each medical symptom into a symptom characteristic of a preset data structure, wherein the symptom characteristic comprises at least one of a symptom part, a symptom parameter and a symptom representation;

and obtaining a word vector of each symptom characteristic according to a mapping relation between a preset symptom entity and the symptom characteristic.

In one possible design, the target symptom entity and the symptom description information are both stored on a blockchain node.

In a second aspect, an embodiment of the present application provides a search apparatus having a function of implementing a search method corresponding to the first aspect. The functions can be realized by hardware, and the functions can also be realized by executing corresponding software by hardware. The hardware or software includes one or more modules corresponding to the above functions, which may be software and/or hardware.

In one possible design, the search apparatus includes:

the input/output module is used for acquiring search information;

a processing module for obtaining at least one symptom feature based on the search information; setting weights for all symptom characteristics according to the semantic association degree of the symptom characteristics and symptoms; acquiring target similarity between each symptom characteristic with the set weight and at least one preset symptom entity; determining a target symptom entity from the at least one preset symptom entity according to the target similarity;

the input and output module is further used for outputting the symptom description information corresponding to the target symptom entity.

In one possible design, the processing module is specifically configured to:

In one possible design, the preset symptom entity corresponds to at least one symptom description information; the processing module is specifically configured to:

In one possible design, the processing module is specifically configured to:

outputting symptom recommendation information obtained according to the search information through the input and output module, wherein the symptom recommendation information corresponds to a preset symptom entity;

acquiring operation behavior data of the user on the symptom recommendation information through the input and output module;

In one possible design, the processing module is specifically configured to:

obtaining an operation behavior map according to the operation behavior data;

and determining the target symptom entity according to the similarity.

In one possible design, the processing module is further configured to:

acquiring training data through the input and output module, wherein the training data comprises a plurality of medical corpora;

In yet another aspect, an embodiment of the present application provides a search apparatus, which includes at least one connected processor, a memory and a transceiver, where the memory is used for storing a computer program, and the processor is used for calling the computer program in the memory to execute the method according to the first aspect.

Yet another aspect of the embodiments of the present application provides a computer-readable storage medium, which includes instructions that, when executed on a computer, cause the computer to perform the method of the first aspect.

Compared with the prior art, in the scheme provided by the embodiment of the application, because at least one symptom characteristic is obtained based on the search information, a large number of manual labels are not needed, and the workload is reduced. According to the semantic association degree of the symptom characteristics and the symptoms, weights are respectively set for the symptom characteristics, the target similarity between the symptom characteristics with the weights and at least one preset symptom entity is obtained, and according to the target similarity, the target symptom entity is determined from the at least one preset symptom entity and output. Therefore, the scheme can improve the accuracy of identifying symptom entities and effectively distinguish similar symptoms from non-similar symptoms.

Drawings

FIG. 1a is a schematic diagram of a network topology of a search system according to an embodiment of the present application;

FIG. 1b is a schematic diagram of a network topology of a search system according to an embodiment of the present application;

FIG. 2 is a schematic flow chart of a searching method in an embodiment of the present application;

FIG. 3 is a schematic diagram of a network topology of a search system according to an embodiment of the present application;

FIG. 4a is a schematic diagram of a network topology of a search system according to an embodiment of the present application;

FIG. 4b is a schematic diagram of a network topology of a search system according to an embodiment of the present application;

FIG. 5 is a schematic diagram of a network topology of a search system according to an embodiment of the present application;

FIG. 6 is a schematic diagram of a blockchain system according to an embodiment of the present application;

FIG. 7 is a schematic diagram of a searching apparatus according to an embodiment of the present application;

FIG. 8 is a schematic diagram of a searching apparatus according to an embodiment of the present application;

fig. 9 is a schematic structural diagram of a server in an embodiment of the present application.

Detailed Description

The terms "first," "second," and the like in the description and in the claims of the embodiments of the application and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It will be appreciated that the data so used may be interchanged under appropriate circumstances such that the embodiments described herein may be practiced otherwise than as specifically illustrated or described herein. Furthermore, the terms "comprise" and "have," and any variations thereof, are intended to cover non-exclusive inclusions, such that a process, method, system, article, or apparatus that comprises a list of steps or modules is not necessarily limited to those steps or modules expressly listed, but may include other steps or modules not expressly listed or inherent to such process, method, article, or apparatus, such that the division of modules presented in the present application is merely a logical division and may be implemented in a practical application in a different manner, such that multiple modules may be combined or integrated into another system or some features may be omitted or not implemented, and such that couplings or direct couplings or communicative connections shown or discussed may be through interfaces, indirect couplings or communicative connections between modules may be electrical or the like, the embodiments of the present application are not limited. Moreover, the modules or sub-modules described as separate components may or may not be physically separated, may or may not be physical modules, or may be distributed in a plurality of circuit modules, and some or all of the modules may be selected according to actual needs to achieve the purpose of the embodiments of the present application.

The embodiment of the application provides a searching method, a searching device and a storage medium, which can be used for a search engine, and the search engine side can be used for providing operations such as inquiry of medical related knowledge, department guidance and the like for a user. The scheme can be used for a server side or a user equipment side, in the embodiment of the application, only a search engine is deployed on the server side as an example, and a search device is deployed on the server side.

The scheme provided by the embodiment of the present application relates to technologies such as Artificial Intelligence (AI), Natural Language Processing (NLP), Machine Learning (ML), and specifically is described by the following embodiments:

the AI is a theory, method, technique and application system that simulates, extends and expands human intelligence, senses the environment, acquires knowledge and uses the knowledge to obtain the best results using a digital computer or a machine controlled by a digital computer. In other words, artificial intelligence is a comprehensive technique of computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can react in a manner similar to human intelligence. Artificial intelligence is the research of the design principle and the realization method of various intelligent machines, so that the machines have the functions of perception, reasoning and decision making.

The AI technology is a comprehensive subject, and relates to the field of extensive technology, both hardware level technology and software level technology. The artificial intelligence infrastructure generally includes technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technologies, operation/interaction systems, mechatronics, and the like. The artificial intelligence software technology mainly comprises a computer vision technology, a voice processing technology, a natural language processing technology, machine learning/deep learning and the like.

NLP is an important direction in the fields of computer science and artificial intelligence. It studies various theories and methods that enable efficient communication between humans and computers using natural language. Natural language processing is a science integrating linguistics, computer science and mathematics. Therefore, the research in this field will involve natural language, i.e. the language that people use everyday, so it is closely related to the research of linguistics. Natural language processing techniques typically include text processing, semantic understanding, machine translation, robotic question and answer, knowledge mapping, and the like.

ML is a multi-field interdisciplinary, and relates to a plurality of disciplines such as probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory and the like. The special research on how a computer simulates or realizes the learning behavior of human beings so as to acquire new knowledge or skills and reorganize the existing knowledge structure to continuously improve the performance of the computer. Machine learning is the core of artificial intelligence, is the fundamental approach for computers to have intelligence, and is applied to all fields of artificial intelligence. Machine learning and deep learning generally include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and formal education learning.

In some embodiments, a search system as shown in FIG. 1a includes a search engine, at least one user terminal. The user inputs search information on a front-end page of a search engine through the user terminal, after the search engine obtains the search information, the search information can be segmented, symptom features which the user wants to describe are obtained through analysis, and then entity mapping is carried out on the symptom features on the basis of a preset knowledge base, so that the actual symptoms of the user are obtained, and a clear basis is provided for subsequent symptom judgment and department guidance.

Entity mapping: the entity mapping refers to mapping entity mentions in the text to a preset knowledge base, and plays a very interesting basic role in the fields of question answering, semantic searching, information extraction and the like. Symptom entity mapping in the medical field is indispensable in medical intelligent assistant application, and for example, stomachache/lower abdomen pain and the like need to be linked to a standard entity word of abdominal pain first and then be subjected to subsequent processing. Entity mapping may also be referred to as entity linking, which is not limited in the embodiments of the present application.

Similarity: the similarity is a quantitative measurement method for measuring the character face similarity of two character strings. The principle is to change one character string into another character string through certain operation. The fewer steps are required, the smaller the distance between the two character strings is, and the greater the similarity of the character strings is. The similarity is a levenstein distance, and comprises deletion, insertion and replacement and other operations. Of course, there are other similarity definitions such as edit distance, Longest Common Subsequence (LCS), and hamming distance, which are not limited in the embodiments of the present application.

Weighted similarity: the likelihood of deletion is different due to different instances of insertion, replacement, and deletion. For example, when spelling correction is performed, letters in some positions are more likely to be mistaken for a certain letter than other letters; some bases in the gene sequence are more likely to be deleted and substituted than others. On the basis of the similarity, adding different weights for deletion, insertion, replacement and other operations is called weighted similarity.

Presetting a knowledge base: refers to a knowledge base for providing entity traversal and entity mapping. In some embodiments, the predetermined knowledge base may be implemented by a neural network model, specifically, a large amount of medical corpus in the medical field is obtained first, then a word vector model is trained based on the medical corpus, and a symptom mapping relationship (i.e., a predetermined knowledge base) is constructed based on various basic symptoms in the medical field. The symptom mapping relation comprises mapping relations among a plurality of preset symptom entities, symptom parts, symptom degrees and symptom representations. And training based on a word vector model to obtain word vectors of the basic symptoms. After the search information input by the user is acquired, the search information is input into the word vector model, and word vectors corresponding to symptom features represented by the search information are acquired. Then, the similarity between the word vector and the word vector corresponding to each preset symptom entity in the symptom mapping relationship is calculated, candidate symptom entities with the similarity higher than a preset threshold are selected from the word vector and the word vector, and one candidate symptom entity can be selected from the candidate symptom entities as a target symptom entity mapped with the search information (for example, the candidate symptom entity with the maximum similarity is selected). In some embodiments, in the present application example, the word vector model may adopt a neural network model structure as shown in fig. 1b, and the word vector model includes an input layer, a hidden layer, and an output layer. The hidden layer is formed by at least two layers of recurrent neural networks, the recurrent neural network in the language model may be a Long Short Term Memory (LSTM) network (fig. 1b takes two layers of LSTM networks as an example), a Gated Recurrent Unit (GRU), a Simple Recurrent Unit (SRU), and other neural networks, which is not limited in the embodiment of the present application. The pre-training process of the word vector model is described as follows: obtaining training data, the training data comprising a plurality of medical corpora (e.g., word features in FIG. 1 b); determining a plurality of medical symptoms (i.e., syntactic features in FIG. 1 b) from the plurality of medical corpuses; respectively converting each medical symptom into a symptom feature (namely, a grammatical feature in fig. 1 b) of a preset data structure, wherein the symptom feature comprises at least one of a symptom part, a symptom parameter and a symptom representation; and obtaining a word vector of each symptom characteristic according to a mapping relation between a preset symptom entity and the symptom characteristic. And training the model based on the word vector of each symptom characteristic to obtain a word vector model. After the word vector model is trained, the features can be extracted from the search information input by the user on line, then the features are integrated, and then the word vector of each feature in the word vector model is obtained.

It should be particularly noted that the server related to the embodiment of the present application may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server providing basic cloud computing services such as a cloud service, a cloud database, cloud computing, a cloud function, cloud storage, a network service, cloud communication, a middleware service, a domain name service, a security service, a CDN, a big data and artificial intelligence platform. The user equipment related to the embodiment of the application may be a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart speaker, a smart watch, and the like, but is not limited thereto. The user equipment and the server may be directly or indirectly connected through wired or wireless communication, and the embodiment of the present application is not limited herein.

Referring to fig. 2, a search method provided in an embodiment of the present application is described below, where the embodiment of the present application includes:

201. and acquiring search information.

The search information is input by the user on a front page of the search engine, and for example, the search information is phrases such as "cold", "shank is somehow through", "leg is very painful", "lower limb feels pain", and the like. The search information may also be referred to as a search word, a keyword, a search phrase, and the like, and the embodiment of the present application does not limit the name and the obtaining manner of the search information.

202. At least one symptom feature is derived based on the search information.

Wherein, the symptom characteristic refers to the state of subjective abnormal feeling or some objective pathological changes of the patient caused by a series of abnormal changes of functions, metabolism and morphological structures in the organism in the disease process. Symptoms are presented. For example, where the search information is encephalitis, the symptom signature may include systemic symptoms: for example, fever, mental fatigue and hypodynamia; symptoms of the whole brain: such as headache, nausea, vomiting, epilepsy, convulsions, coma, disturbance of consciousness; focal symptoms: hemiplegia and aphasia.

In some embodiments, the information obtained by segmenting the search information into the preset format may be used to obtain at least one symptom characteristic; or guiding the user to generate user click behavior, and obtaining at least one symptom characteristic based on the user click behavior. The following are described separately:

the first method is as follows: the method comprises the steps of segmenting search information into information in a preset format.

For the similar entities of the word, because of the diversity of the user input, the search information can be segmented into the information of the preset format, and the preset format can comprise three items: location of symptoms, extent of symptoms, and characterization of symptoms. Even if one or two of the symptom part, the symptom degree and the symptom representation are not included in the search information, the search information after word segmentation can be represented by the preset format, and the embodiment of the application allows a part of the information in the preset format to be represented as empty.

For example, first, the search information (e.g., user query) is segmented into: symptom site + degree of symptom + symptom characterization, each part contains some candidate words. Symptom site words such as abdominal pain include: belly, abdomen, etc., and the common symptom degree words include: some, a little, mild, slight, etc., and common symptoms are characterized by: pain, aches, pains, etc. As shown in fig. 3.

Therefore, the user query is subjected to word segmentation by combining medical treatment relevance, namely the user query is decomposed by using a preset format of symptom part + symptom degree + symptom representation, and a small amount of expressions of different symptoms are formulated, so that the problems that complete listing cannot be realized by using pure mapping table mapping and a large amount of manual labeling required by a supervision model in the prior art can be solved.

The second method comprises the following steps: the method comprises the steps of generating a user click behavior by guiding a user, and obtaining at least one symptom characteristic based on the user click behavior.

In the second method, considering the diversity and uncertainty of the search information input by the user, even if the user describes the same thing, due to the corpus limitation in the search engine, the similarity between the search information input by the user and the corpus in the search engine may be low in terms of characters or semantics, and a situation that the preset symptom entity cannot be matched may occur. For example, when the search information is "not asleep at night", the preset symptom entity may be "insomnia", or when the search information is "toilet many times a day", the preset symptom entity may be "frequent urination". Because the languages are completely different literally, and the related linguistic data are difficult to obtain, so that the training of semantic vectors is difficult, the entity mapping is difficult to be carried out from static semantics, namely the entity mapping is difficult to be carried out on 'sleeping at night' and 'insomnia', and the entity mapping is difficult to be carried out on 'toilet for a plurality of times a day' and 'frequent micturition'.

In order to avoid the situation and improve the effective response to the user, the embodiment of the application may further perform entity mapping for the symptom expressed by the explicit search information based on the user click behavior, for example, based on a dynamic click log of the user, the click data path is strongly associated, and the relevant candidate symptom entity may be found out by selecting and filtering a certain threshold. Specifically, the obtaining at least one symptom feature based on the search information includes:

The symptom recommendation information is preset by a search engine, and can be terms, articles, information, questions and answers and other contents.

For example, when the search information is "not asleep at night", the symptom recommendation information may include: falling asleep at 3 o' clock every day (affecting health), waking up suddenly often in the middle of the night and then falling asleep (aging); or when the search information is "toilet multiple times a day", the symptom recommendation information may include: the patient can go to the toilet for many times within 1 hour (counting the frequency of urination) and frequently go to the toilet when drinking a little water (the bladder has a problem). For another example, when the search information input by the user is "cold", as shown in fig. 4a, the search engine may transmit the following symptom recommendation information to the user terminal: cough (pneumonia), 38.4 degrees (fever with cold), high fever (cold water), fever (how to do), headache (how to do), and the like, and then click behaviors of the user recommending information for the symptoms are detected, and it is assumed that the user clicks "38.4 degrees (fever with cold)" for 8 times, clicks "high fever (cold water with cold water)" for 6 times, and clicks "fever (how to do)" for 13 times. Then, the user click behavior data can be collected, then statistical analysis is performed based on the user click behavior data to obtain a symptom relationship map as shown in fig. 4b, and finally, a target symptom entity matched with the search information of the user is determined, so that the entity mapping operation to the preset knowledge base is facilitated.

203. And respectively setting weights for the symptom characteristics according to the semantic association degree of the symptom characteristics and the symptoms.

The semantic relevance refers to the relevance between a word in the search information and a symptom of a certain disease when the word is used for representing the symptom.

A weight refers to the degree of importance of a factor or indicator relative to an event. For example, a weight may refer to the importance of a certain word in the search information for characterizing a certain symptom of a condition.

In some embodiments, since the medically related words, such as the symptom parts, the symptom representations, and other words, can more effectively distinguish different symptoms, the weights of the symptom parts and the symptom representations, which can effectively distinguish different symptoms, should be as large as possible, and the weights of the other words, such as the words of the symptom degrees, the time when the symptoms occur, and the like, are not high for the degree of distinguishing the symptoms, and should be as small as possible. Based on this, the embodiment of the present application proposes an entity mapping method based on weighted edit distances of medical parts, symptoms and symptom degrees and an alias entity mapping library to solve the problem. Specifically, the setting of the weight for each symptom feature according to the semantic relevance between the symptom feature and the symptom includes:

It can be seen that in some embodiments, if the simple similarity is used to find the similarity of the query of the user, erroneous judgment is likely to occur, for example, if the user inputs that "the stomach is itchy" and the word is segmented, and then obtains that "the stomach is itchy" and if the simple editing distance is used, the editing distance of "the leg is itchy" in the leg pain entity is smaller and closer, and the query of the user is likely to be incorrectly linked to the leg pain entity. Also if the user inputs "light and gentle calf pain" this entity of abdominal pain is misjudged for the same reason. Therefore, the embodiment of the application has different importance for the medical symptoms, and gives larger weight to the part and symptom characterization words, and gives smaller weight to the degree word and other medical irrelevant words. For example, the symptom part word and the symptom characterizing word are respectively given different weights of more than 2, the symptom degree adverb is given a weight of more than 1 and less than 2, and other irrelevant words are given a weight of 1. And then, calculating the weighted editing distances of different expressions of different preset symptom entities in the input and formulated entity mapping table of the user and calculating the average value to obtain the average weighted editing distance between the search information and each preset symptom entity. And the preset symptom entity with the minimum average edit distance is regarded as the symptom entity closest to the search information. As shown in fig. 3, four different user queries are weighted and edited with different expressions in the two entities of predetermined symptoms of abdominal pain and leg pain, respectively, to obtain an average value. It can be seen that "the stomach is a bit painful" and "the lower leg is a bit itchy" can be successfully mapped to two pre-set symptom entities "abdominal pain" and "leg itch", respectively.

Therefore, by respectively giving different weights to three items in the information in the preset format, the symptoms which the user table actually wants to express can be reflected more accurately, meanwhile, the symptoms of the user are more definite, and the condition that the search information of the user is mapped to the wrong preset symptom entity when the similarity is calculated subsequently is avoided.

204. And acquiring target similarity between each symptom characteristic with the set weight and at least one preset symptom entity.

Wherein, the preset symptom entity refers to an entity for characterizing the symptom characteristics of a certain disease. For example, when the symptom is leg pain, lower limb pain, and the like, the predetermined symptom entity may include a leg pain entity; when the symptom is headache, headache with some pain and the like, the preset symptom entity can comprise a headache entity; when the symptom is abdominal pain, abdominal discomfort and the like, the preset symptom entity can comprise an abdominal pain entity; when the symptom is sleeplessness, insomnia, or the like, the predetermined symptom entity may include a poor sleep entity. And so on, will not be described in detail. Specifically, reference may be made to the preset symptom entity shown in fig. 3.

In some embodiments, the preset symptom entity corresponds to at least one symptom description information, and the symptom description information refers to information describing the characteristics of the symptom caused by the body and the spirit. Considering the diversity of the search information input by the user, when entity mapping is performed, one search information may be mapped to a plurality of preset symptom entities, and the preset symptom entities that may be obtained do not completely and definitely reflect the search requirement of the user, then a more reasonable target similarity may be obtained by using a weighting method, specifically, the obtaining of the target similarity between each symptom feature and the preset symptom entity includes:

For example, the symptom part word and the symptom characterizing word are respectively given different weights of more than 2, the symptom degree adverb is given a weight of more than 1 and less than 2, and other irrelevant words are given a weight of 1. And then, calculating the weighted editing distances of different expressions of different preset symptom entities in the input and formulated entity mapping table of the user and calculating the average value to obtain the average weighted editing distance between the search information and each preset symptom entity. And the preset symptom entity with the minimum average edit distance is regarded as the symptom entity closest to the search information. As shown in fig. 3, four different user queries are weighted and edited with different expressions in the two entities of predetermined symptoms of abdominal pain and leg pain, respectively, to obtain an average value. It can be seen that "the stomach is a bit painful" and "the lower leg is a bit itchy" can be successfully mapped to two pre-set symptom entities "abdominal pain" and "leg itch", respectively.

It can be seen that, by weighting the similarity measure, the search information of the user can be mapped to the wrong preset symptom entity.

205. And determining a target symptom entity from the at least one preset symptom entity according to the target similarity and outputting the target symptom entity.

In some embodiments, when the at least one symptom feature is obtained in the second manner, the operation behavior map corresponding to the user may be further analyzed based on the operation behavior data, so as to identify the target symptom entity. Specifically, the determining a target symptom entity from the at least one preset symptom entity according to the target similarity includes:

obtaining an operation behavior map according to the operation behavior data;

and determining the target symptom entity according to the similarity.

The operation behavior map can represent click paths, click objects, click times, click sequence, click time intervals and the like of the user for symptom recommendation information, and can reflect multi-dimensional information of symptoms which the user wants to express for the search information.

In some embodiments, based on the entity mapping part in the basic symptom representation list shown in fig. 3, when a user inputs a symptom entity a (e.g., the aforementioned search information, user query), as shown in fig. 5, first, candidate symptom entities are found based on the weighted edit distance measurement model, and if the calculated similarity a is greater than a preset threshold, the symptom entity a may be linked to a preset symptom entity a in the basic symptom representation list that is most relevant to the symptom entity a.

If the similarity a is greater than or equal to the predetermined threshold, the similarity b between the symptom entity a and the predetermined symptom entity a can be calculated by using the word vector model (e.g., the word vector model shown in fig. 1 b). Specifically, if the calculated similarity b is greater than a preset threshold, entity linking is performed in the same way; if the similarity b is smaller than or equal to a preset threshold, generating candidate symptom entities according to the click behavior data of the user, then calculating the similarity c between the candidate symptom entities and each preset symptom entity in the basic symptom representation list, and if the similarity c is larger than the preset threshold, performing entity linking.

In the embodiment, through the three steps, the accuracy and the recall rate of the symptom entity link can be guaranteed to the greatest extent, and the method has strong practicability.

In the embodiment of the application, on one hand, because at least one symptom feature is obtained based on the search information, a large number of manual labels are not needed, and thus the workload is reduced. According to the semantic association degree of the symptom characteristics and the symptoms, weights are respectively set for the symptom characteristics, the target similarity between the symptom characteristics with the weights and at least one preset symptom entity is obtained, and according to the target similarity, the target symptom entity is determined from the at least one preset symptom entity and output. Therefore, the scheme can improve the accuracy of identifying symptom entities and effectively distinguish similar symptoms from non-similar symptoms.

On the other hand, by respectively giving different weights to three items in the information in the preset format, the symptom which the user table actually wants to express can be more accurately reflected, meanwhile, the symptom of the user is more definite, and the condition that the search information of the user is mapped to the wrong preset symptom entity when the similarity is calculated subsequently is avoided.

In the embodiment of the present application, the target symptom entity and the symptom description information may be both stored in the blockchain. The blockchain is a novel application mode of computer technologies such as distributed data storage, point-to-point transmission, a consensus mechanism and an encryption algorithm. A block chain (Blockchain), which is essentially a decentralized database, is a series of data blocks associated by using a cryptographic method, and each data block contains information of a batch of network transactions, so as to verify the validity (anti-counterfeiting) of the information and generate a next block. The blockchain may include a blockchain underlying platform, a platform product services layer, and an application services layer.

The block chain underlying platform can comprise processing modules such as user management, basic service, intelligent contract and operation monitoring. The user management module is responsible for identity information management of all blockchain participants, and comprises public and private key generation maintenance (account management), key management, user real identity and blockchain address corresponding relation maintenance (authority management) and the like, and under the authorization condition, the user management module supervises and audits the transaction condition of certain real identities and provides rule configuration (wind control audit) of risk control; the basic service module is deployed on all block chain node equipment and used for verifying the validity of the service request, recording the service request to storage after consensus on the valid request is completed, for a new service request, the basic service firstly performs interface adaptation analysis and authentication processing (interface adaptation), then encrypts service information (consensus management) through a consensus algorithm, transmits the service information to a shared account (network communication) completely and consistently after encryption, and performs recording and storage; the intelligent contract module is responsible for registering and issuing contracts, triggering the contracts and executing the contracts, developers can define contract logics through a certain programming language, issue the contract logics to a block chain (contract registration), call keys or other event triggering and executing according to the logics of contract clauses, complete the contract logics and simultaneously provide the function of upgrading and canceling the contracts; the operation monitoring module is mainly responsible for deployment, configuration modification, contract setting, cloud adaptation in the product release process and visual output of real-time states in product operation, such as: alarm, monitoring network conditions, monitoring node equipment health status, and the like.

The searching apparatus (also referred to as a server) performing the searching method in the embodiment of the present application may be a node in a blockchain system. The searching apparatus in the embodiment of the present application may be a node in a block chain system as shown in fig. 6.

Any technical feature mentioned in the embodiment corresponding to any one of fig. 1a to 6 is also applicable to the embodiment corresponding to fig. 7 to 9 in the embodiment of the present application, and the details of the subsequent similarities are not repeated.

A search method in the embodiment of the present application is described above, and a device for performing the search method is described below.

Referring to fig. 7, a schematic diagram of a search apparatus 70 shown in fig. 7 is applicable to a search engine, which can be used for providing a query of medical-related knowledge to a user, department guidance, and the like. The searching apparatus 70 in the embodiment of the present application can implement the steps corresponding to the searching method performed in the embodiment corresponding to any one of fig. 1a to 5. The functions implemented by the search device 70 may be implemented by hardware, or may be implemented by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above functions, which may be software and/or hardware. The search apparatus 70 may include a processing module 701 and an input/output module 702, and the processing module 701 and the input/output module 702 may refer to operations performed in any one of the embodiments corresponding to fig. 1a to 5 to determine a symptom characteristic, a target similarity, a target symptom entity, and the like, which are not described herein again. For example, the processing module 701 may be used to control the operations of the input/output module 702 such as obtaining, inputting and outputting.

In some embodiments, the input output module 702 may be configured to obtain search information;

the processing module 701 may be configured to derive at least one symptom feature based on the search information; setting weights for all symptom characteristics according to the semantic association degree of the symptom characteristics and symptoms; acquiring target similarity between each symptom characteristic with the set weight and at least one preset symptom entity; determining a target symptom entity from the at least one preset symptom entity according to the target similarity;

the input/output module 701 is further configured to output the symptom description information corresponding to the target symptom entity determined by the processing module 701.

In the embodiment of the present application, the processing module 701 obtains at least one symptom feature based on the search information, so that a large number of manual labels are not relied on, and thus the workload is reduced. According to the semantic association degree of the symptom characteristics and the symptoms, weights are respectively set for the symptom characteristics, the target similarity between the symptom characteristics with the weights and at least one preset symptom entity is obtained, and according to the target similarity, the target symptom entity is determined from the at least one preset symptom entity and output. Therefore, the scheme can improve the accuracy of identifying symptom entities and effectively distinguish similar symptoms from non-similar symptoms.

In some embodiments, the processing module 701 is specifically configured to:

In some embodiments, the predetermined symptom entity corresponds to at least one symptom description information; the processing module 701 is specifically configured to:

In some embodiments, the processing module 701 is specifically configured to:

outputting symptom recommendation information obtained according to the search information through the input/output module 702, wherein the symptom recommendation information corresponds to a preset symptom entity;

acquiring the operation behavior data of the user on the symptom recommendation information through the input and output module 702;

In some embodiments, the processing module 701 is specifically configured to:

obtaining an operation behavior map according to the operation behavior data;

and determining the target symptom entity according to the similarity.

The search apparatus 70 in the embodiment of the present application is described above from the perspective of a modular functional entity, and the servers that execute the search method in the embodiment of the present application are described below from the perspective of hardware processing. It should be noted that the entity device corresponding to the processing module 701 may be a processor, and the entity device corresponding to the input/output module 702 in the embodiment shown in fig. 7 in this application may be an input/output unit, a transceiver, a radio frequency circuit, a communication module, an output interface, and the like. The apparatus 70 shown in fig. 8 may have a structure as shown in fig. 8, when the apparatus 70 shown in fig. 7 has a structure as shown in fig. 8, the processor and the input/output unit in fig. 8 can implement the same or similar functions of the processing module 701 and the input/output module 702 provided in the apparatus embodiment corresponding to the search apparatus 70, and the memory in fig. 8 stores a computer program that the processor needs to call when executing the search method.

Fig. 9 is a schematic diagram of a server 920 according to an embodiment of the present disclosure, where the server 920 may have a relatively large difference due to different configurations or performances, and may include one or more Central Processing Units (CPUs) 922 (e.g., one or more processors) and a memory 932, and one or more storage media 930 (e.g., one or more mass storage devices) for storing applications 942 or data 944. Memory 932 and storage media 930 can be, among other things, transient storage or persistent storage. The program stored on the storage medium 930 may include one or more modules (not shown), each of which may include a series of instruction operations for the server. Still further, a central processor 922 may be provided in communication with storage medium 930 to execute a sequence of instruction operations in storage medium 930 on server 920.

The Server 920 may also include one or more power supplies 926, one or more wired or wireless network interfaces 950, one or more input-output interfaces 959, and/or one or more operating systems 941, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc.

The steps performed by the server in the above embodiment may be based on the structure of the server 920 shown in fig. 9. The steps performed by the apparatus 60 shown in fig. 9 in the above-described embodiment may be based on the server structure shown in fig. 9, for example. For example, the processor 922, by invoking instructions in the memory 932, performs the following:

acquiring search information through the input/output interface 959;

obtaining at least one symptom feature based on the search information; setting weights for all symptom characteristics according to the semantic association degree of the symptom characteristics and symptoms; acquiring target similarity between each symptom characteristic with the set weight and at least one preset symptom entity; determining a target symptom entity from the at least one preset symptom entity according to the target similarity;

and outputting symptom description information corresponding to the target symptom entity determined by the processing module 701 through the input/output interface 959.

In the foregoing embodiments, the descriptions of the respective embodiments have respective emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to related descriptions of other embodiments.

It can be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working processes of the system, the apparatus and the module described above may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.

In the several embodiments provided in the embodiments of the present application, it should be understood that the disclosed system, apparatus, and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the modules is merely a logical division, and in actual implementation, there may be other divisions, for example, multiple modules or components may be combined or integrated into another system, or some features may be omitted, or not implemented. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or modules, and may be in an electrical, mechanical or other form.

The modules described as separate parts may or may not be physically separate, and parts displayed as modules may or may not be physical modules, may be located in one place, or may be distributed on a plurality of network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment.

In addition, functional modules in the embodiments of the present application may be integrated into one processing module, or each module may exist alone physically, or two or more modules are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module, if implemented in the form of a software functional module and sold or used as a stand-alone product, may be stored in a computer readable storage medium.

In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product.

The computer program product includes one or more computer instructions. The procedures or functions described in accordance with the embodiments of the present application are generated in whole or in part when the computer program is loaded and executed on a computer. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, from one website site, computer, server, or data center to another website site, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can store or a data storage device, such as a server, a data center, etc., that is integrated with one or more available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., Solid State Disk (SSD)), among others.

The technical solutions provided by the embodiments of the present application are introduced in detail, and the principles and implementations of the embodiments of the present application are explained by applying specific examples in the embodiments of the present application, and the descriptions of the embodiments are only used to help understanding the method and core ideas of the embodiments of the present application; meanwhile, for a person skilled in the art, according to the idea of the embodiment of the present application, there may be a change in the specific implementation and application scope, and in summary, the content of the present specification should not be construed as a limitation to the embodiment of the present application.

Claims

1. A method of searching, the method comprising:

acquiring search information;

obtaining at least one symptom feature based on the search information;

2. The method according to claim 1, wherein the setting of the weight for each symptom feature according to the semantic relevance between the symptom feature and the symptom comprises:

3. The method according to claim 2, wherein the preset symptom entity corresponds to at least one symptom description information; the obtaining of the target similarity between each symptom feature and a preset symptom entity includes:

4. The method according to any one of claims 1-3, wherein the deriving at least one symptom feature based on the search information comprises:

5. The method of claim 4, wherein the determining a target symptom entity from the at least one preset symptom entity according to the target similarity comprises:

obtaining an operation behavior map according to the operation behavior data;

and determining the target symptom entity according to the similarity.

6. The method of claim 4, further comprising:

7. The method of claim 1, wherein the target symptom entity and the symptom description information are stored on a blockchain node.

8. A search apparatus, characterized in that the search apparatus comprises:

the input/output module is used for acquiring search information;

9. A search apparatus, characterized in that the search apparatus comprises:

at least one processor, memory, and transceiver;

wherein the memory is for storing a computer program and the processor is for calling the computer program stored in the memory to perform the method as claimed in any one of claims 1-7.

10. A computer-readable storage medium characterized in that it comprises instructions which, when run on a computer, cause the computer to perform the method as claimed in any one of claims 1-7.