CN110310739B - Health information processing method and system - Google Patents
Health information processing method and system Download PDFInfo
- Publication number
- CN110310739B CN110310739B CN201810232176.8A CN201810232176A CN110310739B CN 110310739 B CN110310739 B CN 110310739B CN 201810232176 A CN201810232176 A CN 201810232176A CN 110310739 B CN110310739 B CN 110310739B
- Authority
- CN
- China
- Prior art keywords
- information
- input
- output
- output information
- health
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment
Landscapes
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Pathology (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
A health information processing method and system. The method and the system solve the problem of carrying out system processing on the complex information with certainty and/or uncertainty, the system can process various information combinations by inputting the setting and classifying the decision tree, and the output information is obtained by matching the information on the basis of the corresponding database. The invention is used for carrying out high-efficiency personalized system processing on complex information which is difficult to process, and can assist experts in the health field to efficiently provide personalized information for users when the system is specific to the health information field.
Description
Technical Field
The present invention relates to the field of information processing technologies, and in particular, to a method and a system for processing health information.
Background
Efficient processing of complex information is an important means for facilitating the circulation of complex information. However, in some systems, such as the medical health field, the types of related information are complicated and various, the related information is scattered, and the professional supply is insufficient, so that the efficiency of manually processing the mixed type complex information is low. Furthermore, since the following may exist between the input information and the output information: coexistence of specific relation and non-specific relation; several input sets with definite and uncertain relations to the output information coexist; the input information is not independent; the output information is not independent; there is an unknown mechanism; data sets are difficult to acquire, etc., thereby making conventional approaches unsuitable for systematic processing of such complex information.
The current common schemes for processing mixed complex information (miscellaneous information) include consulting services, popular science articles, scientific literature and the like provided by professionals.
Among them, the counseling service provided by professionals is a way of processing complex information, with similar functions of giving corresponding outputs according to inputs. In this solution, the professional can provide feedback with certain integration to some extent, but this solution has the disadvantages of limited supply capacity, high cost, time consuming, etc. In addition, the objectivity and knowledge background of the information provider can limit the quality and content of the feedback.
The science popularization article has features that are easy to understand. However, the quality depends on the objectivity and professional level of the author, some of the quality is based on the promotion of products, services and the like, and potential interest conflict can result in lack of objectivity and needs to be carefully screened by information demanders. In addition, although the science popularization article can comprehensively record related information of a certain aspect, the content is mainly based on an information algorithm, the integration degree is low, an information demand party may need to enlarge reading, and effective information can be extracted by the information demand party.
Scientific literature generally aims at a certain target, collects relevant data, analyzes the data through a proper analysis method to obtain a conclusion, and transmits the conclusion in a scientific literature mode. The disadvantage of this scheme is that the content is relatively obscure and the requirement for audience is relatively high; in addition, only limited targets are targeted, lacking in systematicness, and system information needs to be provided along with other relevant research. In addition, most of the analyses are based on available data, and the analysis method is limited by certain preconditions, which influence the universality and effectiveness of the application.
In a word, the above schemes all have small supply capacity, low efficiency and high cost, and cannot meet the requirements of most people in real life on complex information system processing, especially the acquisition requirements of many people in the medical health field on early, middle and late cautions and related information of diseases.
Disclosure of Invention
In view of the above, the present invention is directed to a method and system for processing health information, which at least partially solve at least one of the above problems.
In order to achieve the above object, as one aspect of the present invention, there is provided a health information processing method including the steps of:
step 2, matching the input information of different categories with corresponding output information according to different input combinations obtained by the decision tree classification in the step 1, and storing the input information as an intermediate result; wherein the output information can also be in the form of a set;
step 3, selecting a corresponding processing mode for the intermediate result according to different input combinations obtained by the decision tree classification in the step 1 to obtain corresponding final output information;
the health information expert system comprises a health information expert system, a health information database and a health information database, wherein the input information and the output information are information related to health, and the health information database stores various input and output corresponding relations;
the health information database can define primary outputs and primary output subsets according to requirements in a construction process.
As another aspect of the present invention, there is provided a health information processing system characterized by comprising:
a memory storing a program for executing the health information processing method as described above;
a processor for executing the program stored in the memory.
The health information processing system also comprises a plurality of information databases, and the information databases store various input and output corresponding relations;
the information database can define primary output and primary output subsets according to requirements in the construction process.
Compared with the prior art, the technical scheme of the invention has the following beneficial effects: the process from complex input to effective personalized information output can be efficiently completed, and the supply capacity is large; only the interaction between the user and the system is involved in the use process, so that the man-made interference caused by the reasons of background, professional morality, benefit conflict and the like in the information transmission is avoided; the system has good expansibility, and can efficiently provide information at any time, so that the scheme has great superiority in coverage, circulation time and efficiency; the invention realizes the real-time, objective, economic and efficient supply of effective personalized information under the condition of limited conditions, establishes a system for efficiently processing complex information and solves the problem of unsmooth circulation of the complex information; when the method is specifically applied to medical information, the user can be given basic knowledge of prevention and attention such as relevant reference information and attention items at any period of diseases without the help of professional medical practitioners, and the user can be helped and promoted to improve the self state.
Drawings
FIG. 1 is a schematic diagram of specific inputs, non-specific inputs, and specific/non-specific mixed inputs;
FIG. 2 is a schematic diagram of a classification decision tree for each input message of the system (illustrated with 3 inputs);
FIG. 3 is a flow chart of the sorting and ordering process of the present invention;
FIG. 4 is a flow diagram illustrating output subset processing for non-zero, indeterminate state output information in an embodiment of the invention;
FIG. 5 illustrates a plug-in configuration in one embodiment of the invention;
fig. 6 is a flow chart of a hybrid health information processing method of the present invention.
Detailed Description
In order that the objects, technical solutions and advantages of the present invention will become more apparent, the present invention will be further described in detail with reference to the accompanying drawings in conjunction with the following specific embodiments.
The invention provides a mixed complex information (mixed information) for complex information, especially for complex information with a specific relation and a non-specific relation, a plurality of input sets with a definite relation and an uncertain relation with output information, non-independence between input information, non-independence between output information, an unknown mechanism, difficult acquisition of data sets and the like. The output may be multi-level depending on the requirements. The system can obtain corresponding output information step by step according to the primary output information so as to meet the requirements of users on the corresponding information.
It is well known that things are related to each other. The occurrence of an event may be deterministic or non-deterministic. And deducing that the output must occur through the input, wherein the input information is the determined state input information. And deducing that the output may or may not occur through the input, wherein the input information is the uncertain state input information. For example, if the input information is "heart disease", that is, it indicates that the heart disease has been diagnosed, it can be inferred that the heart disease must occur; for example, the input information is "heartache", although there are many possibilities of the related pathological changes, it cannot be inferred whether the pathological changes occur or not; the same is true in other fields, for example, if "rain" is input, the current weather can be inferred to be determined as "rainy day", and if "dense cloud" is input, the weather cannot be inferred to be rainy day, and may be a weather caused by cloudy days, dust emission, or the like.
The definitions for the above partial nouns are given in the following table.
The uncertain state input information may have the problem of unclear boundaries in the classification process. This uncertainty causes a non-specific relationship between input and output, i.e., there may be a plurality of different output information corresponding to the input information, resulting in the generation of noise during information processing. For example, the input information "heartache" mentioned above is a symptom that occurs in many kinds of lesions, and it is difficult to infer a definite lesion, so that there is a possibility that noise exists in the output information.
Due to the complexity of processing information in the information processing process caused by the simultaneous existence of certainty/uncertainty and specificity/non-specificity, the conventional method of searching and sorting in the database only according to keywords obtains results which are poor in gradation and high in noise, and output is not ideal. When the system inputs include a deterministic state input, an indeterminate state input that contributes to noise development, and an indeterminate state input that does not contribute to noise development, the system can provide a relatively comprehensive information processing for the user. In addition, system application scenarios become more extensive.
According to the corresponding relation between the input information and the output information in the system, a series of corresponding connection frames can be drawn to show the relation between the input and the output. Each linking frame can be a specific linking frame, a non-specific linking frame, or a hybrid linking frame, and the corresponding linkage is schematically shown in FIG. 1, where I represents input and O represents output. The invention stores all the corresponding relation information as a series of databases to support the data matching step. Further, in some embodiments, there is a corresponding input database for the inputs and a corresponding output database for the outputs.
The input information and the output information of the system can be single-value input, but preferably adopt a set mode to meet various complex conditions. The input information may comprise a plurality of components. The number of elements may vary between each input component. Each input information is divided into a zero set, an error input, and a non-zero set according to its content. Here, a null set is defined as a set that contains only one element and is 0, indicating no information amount, and a non-null set is defined as a set that contains at least one element and all elements are not 0, indicating that valid information is contained. If a set includes both 0 and other elements, the set is defined as an erroneous input. The system start of each input is set to zero set. This input arrangement allows the system to cover a variety of input combinations. For example, when the system sets 3 input options, input 1, input 2, and input 3, respectively, input 1 may have n options, and input 1 may have 2nSeed combination; input 2 may have m options, then input 2 may have 2mSeed combination; input 3 may have k options, then input 3 may have 2kAnd (4) combination. Thus, the entire system may satisfy 2n+m+kAnd (4) combination. Of course, the real-life situation is less complex. When the database corresponding to each input is large enough, the system can process enough scenes. In summary, the system input setting enables the system to meet various user requirements.
The input in the present application can be selected from preset options, or input through a keyboard or voice. If the information is the preset option, the system needs to be considered as comprehensive as possible during design, and the situation that some information is input and cannot be input is avoided. If the input is direct voice or keyboard input, the input information may have deviation due to problems such as accent or keyboard misoperation, and further verification and error correction are needed.
Preferably, the preset options can be selected by a drop-down list or by a single-selection or multi-selection preset grouping option. Wherein, for the case that some options are particularly numerous, grouping input and/or association input can be carried out. The system can set some input information related to the search information in advance for selection according to needs.
Different from the conventional mode of obtaining nodes through various algorithms, the invention establishes a classification decision tree for the nodes according to the contained contents of each input message, namely a zero set, an error input and a non-zero set, so as to integrate the input messages with complex and scattered relations into a system and enable the definition and output under various conditions to be possible. In addition, the classification decision tree classifies input combinations. In each combination, the system can call a corresponding database according to the input information, define corresponding intermediate parameters, match corresponding output information and define primary output information. The corresponding intermediate parameters provide a basis for the later noise reduction treatment. Since the system input design defines the certainty, non-certainty of the input information, the system can perform individual and/or mixed processing of the certainty information, non-certainty information. The design enables the system to quickly complete information processing aiming at various input information combinations and has strong interpretability.
Since there is a mixed possibility of specific and non-specific correspondences in some input/output paths, there is a possibility of noise interference in the output information of uncertain states. Due to the lack of effective data support and the unknown partial mechanism, the complete noise removal is not practical. In the output result, output results from different sources exist in a mixed mode. Due to the existence of multiple combinations, the combinations may have relationships of independence, inclusion, intersection, equivalence and the like, and interference must be removed first to remove noise, and noise sources are traced back to find characteristics corresponding to target output. According to the invention, the interference is stripped layer by layer, and the target output is sequenced according to the corresponding input frequency output in the interference-free uncertain state, so that part of noise is removed from the uncertain state output, and the purpose of improving the output quality is achieved.
Specifically, the invention discloses a health information processing method, which is characterized by comprising the following steps:
step 2, matching the input information of different categories with corresponding output information according to different input combinations obtained by the decision tree classification in the step 1, and storing the input information as an intermediate result; wherein the output information can also be in the form of a set;
step 3, selecting a corresponding processing mode for the intermediate result according to different input combinations obtained by the decision tree classification in the step 1 to obtain corresponding final output information;
the health information expert system comprises a health information expert system, a health information database and a health information database, wherein the input information and the output information are information related to health, and the health information database stores various input and output corresponding relations;
the health information database can define primary outputs and primary output subsets according to requirements in a construction process.
The decision tree classification step in step 1 specifically includes:
defining input information with certainty as first input information, wherein output information obtained by matching the first input information is first output information;
defining input information which has no determinacy and does not contribute to noise formation as second input information, wherein output information obtained by matching the second input information is second output information;
defining input information which has no determinacy and contributes to noise formation as third input information, wherein output information obtained by matching the third input information is third output information; the second output information and the third output information are jointly called uncertain state output information;
and (4) carrying out decision tree classification by taking whether the first input information, the second input information and the third input information are zero sets, non-zero sets and error inputs as nodes.
Wherein, the different input combinations obtained by the decision tree classification in the step 1 are obtained based on the following rules:
when all the input information is 0, setting a state parameter index and assigning the state parameter index to be 0;
when all the input information is not 0 at the same time and the first input information comprises 0, setting a state parameter index and assigning the state parameter index as 1;
when all input information is not 0 at the same time and the first input information does not include 0, setting a state parameter index and assigning a value of 2;
thus, the step 3 specifically includes:
step 301, judging whether the state parameter index is not 0;
if the judgment result is negative, the input error is indicated or the input information can not generate effective feedback, the system outputs feedback information, and the whole method flow is ended;
step 302, judging whether the state parameter index is not 1;
if the judgment result is yes, go to step 305;
step 303, judging whether the output information of the uncertain state is not null; if the judgment result is negative, the existing input information can not provide further information, and the whole method flow is ended;
and step 304, simultaneously or selectively executing at least one of the following steps through different influences of the input information on the sorting process: firstly, data matching, defining subsequent output step by step, classifying and sequencing the output information of the uncertain state, and thirdly, defining and sequencing an output subset through data matching on the basis of a relevant database; after the execution is finished, outputting feedback information, and finishing the whole process;
step 305, judging whether the output information of the uncertain state is not null;
if the judgment result is negative, defining subsequent output step by step through data matching; outputting a feedback result and finishing the whole method flow;
step 306, judging whether the first output information has an inclusion or repetition relationship with the content in the uncertain state output information as an intermediate result;
if not, go to step 309;
if yes, deleting the content with the inclusion or repetition relation in the uncertain state output information;
step 307, classifying the contents with mutual competition relationship in the second output information and the third output information, and further deleting the contents with mutual competition relationship and repetition from the uncertain state output information of the deleted contents with or repetition relationship;
step 308, judging whether the output information of the uncertain state obtained by the step 307 is not null;
if the judgment result is negative, defining subsequent output step by step through data matching; outputting a feedback result, and ending the whole method flow;
step 309, by inputting different influences of the information on the sorting process, at least one of the following steps is performed simultaneously or selectively: firstly, data matching, defining subsequent output step by step, classifying and sequencing the output information of the uncertain state, and thirdly, defining and sequencing an output subset through data matching on the basis of a relevant database; after the execution of the first step and the third step is finished, feedback information is output, and the whole method flow is finished;
step 310, outputting all feedback information, and ending the whole method flow;
preferably, in steps 304 and 309, the step of performing output subset processing on the non-zero uncertain state output information specifically includes:
step 3041, the system reads in non-zero uncertain state output information;
step 3042, determining whether the uncertain state input information contributing to noise formation is not zero;
if the judgment is no, the further processing is terminated;
if yes, carrying out data matching on the basis of the related database;
step 3043, determining whether the output subsets of the corresponding uncertain states can be sorted;
if not, go to step 3045;
step 3044, sorting the corresponding uncertain state outputs;
step 3045, define the relevant information of the output subset, and end the whole process.
When the judgment result is that the input is wrong or the input information cannot generate effective feedback, the existing information or the preset corresponding feedback information is directly output.
Wherein, the step 310 specifically comprises the following steps:
step 3101, according to the difference of the characteristics of the second output information, determining whether each second output information contributes to noise formation and participates in the frequency calculation related to the ranking;
if the second output information and the third output information both contribute to noise formation and the results can be combined, judging whether the results are negative; go to step 3103;
step 3102, classifying based on the relation between the uncertain state output information matched with different input information, eliminating the uncertain state output information of interference ordering, and obtaining redefined uncertain state output information;
step 3103, calculating the frequency corresponding to each uncertain state output information, and determining whether the non-interfering uncertain state output information can be sorted;
if the effective output information of the uncertain state only contains one element, or contains a plurality of elements but the corresponding frequency of each element is the same, judging whether the result is negative, outputting the relevant information without further data processing, and ending the whole process;
and 3104, sorting the interference-free uncertain state output information according to the frequency, and outputting the related information.
Wherein step 3101 is determined by the characteristics of the selected input; the expression is directly simplified according to the logical relation between the self characteristics of the support data and the support data, or a built-in parameter is defined to determine the path.
The input information is selected from preset options or input through a keyboard or voice;
preferably, the preset options are selected by a drop-down list or a single-selection or multi-selection preset grouping option;
preferably, the preset options are grouped according to the intrinsic rules thereof, and the association function of the options is supported; the association function means that when a keyword is input, the system associates a plurality of keywords related to the keyword according to the keyword for the user to select.
Preferably, the health information processing method further comprises a step 4 of inputting a series of input information based on a time axis in addition, and outputting comprehensive output information based on the time axis by circularly executing the steps 1 to 3, wherein the original output of the steps 1 to 3 is shielded; more preferably, the health information processing method further includes a step 5 of outputting the obtained integrated output information as dynamic output information on a time axis in the form of a graph or animation. Therefore, the invention can more fully display the result and improve the identification capability of the system by processing the knowledge base and introducing the time concept.
The invention also discloses a health information processing system, which is characterized by comprising:
a processor for executing programs stored in the memory;
the health information expert system comprises a health information expert system, a health information expert system and a health information management system, wherein the input information and the output information are information related to health, the health information expert system further comprises a plurality of health information databases, and various input and output corresponding relations are stored in the health information databases.
Preferably, the health information database is capable of defining primary outputs and subsets of primary outputs on demand during a construction process.
The processor is a processor in a singlechip, a desktop, a tablet personal computer, a notebook computer, a computer network and a programmable logic controller;
the technical solution of the present invention is further illustrated by some specific embodiments.
FIG. 2 is a diagram of an embodiment of an input information classification decision tree according to the present invention, in which a preferred embodiment of the present invention is illustrated with three input components as an example. The inputs include input 1, input 2, input 3. The input 1 and the output are defined as definite state input for definite relation, and can be selected in the corresponding input information file. The input 2, the input 3 and the output are in uncertain relation, defined as uncertain state input, and can be selected from corresponding input information files which are the same as or different from the input 1. Input 2 and input 3 have different characteristics and different mechanisms of action from output to output. Such an arrangement contains as many possibilities as possible of input, so that the applicability of the solution of the invention is relatively broad.
The specific flow of the classification decision tree is as follows:
If not, go to step 9.
And step 2, judging whether the input 1 only contains 0.
If the result is no, the information is not further processed for an erroneous input.
And step 3, judging whether the input 2 contains 0.
If not, go to step 7.
And 4, judging whether the input 2 only contains 0.
If the result is no, the information is not further processed for an erroneous input.
If the result is no, the combination is defined as B, i.e. input 1 is 0, input 2 is 0, input 3 is not 0, state parameter index is defined, uncertain state output information is defined by information corresponding to input 3, and various subsequent output information and prompt information are defined by uncertain state output information.
Step 6, determine if input 3 contains only 0.
If the result is yes, the combination is defined as A, namely input 1 is 0, input 2 is 0, input 3 is 0, the state parameter index is defined, and prompt information is defined;
if the result is no, the information is not further processed for an erroneous input.
Step 7, judge whether input 3 contains 0.
If the result is negative, the combination is defined as D, namely the input 1 is 0, the input 2 is not 0, the input 3 is not 0, the state parameter index is defined, the uncertain state output is defined through the information corresponding to the input 2 and the input 3, the uncertain state output is redefined through merging, and various subsequent output information is defined through the uncertain state output to define the prompt information.
Step 8, determine if input 3 contains only 0.
If the result is yes, the combination is defined as C, namely input 1 is 0, input 2 is not 0, input 3 is 0, a state parameter index is defined, uncertain state output is defined through information corresponding to input 2, various subsequent output information is defined through uncertain state output, and prompt information is defined;
if the result is no, the information is not further processed for an erroneous input.
Step 9, determine if input 2 contains 0.
If not, go to step 13.
Step 10, determine if input 2 contains only 0.
If the result is no, the information is not further processed for an erroneous input.
Step 11, determine if input 3 contains 0.
And if the result is negative, defining the combination as F, namely the input 1 is not 0, the input 2 is 0, the input 3 is not 0, defining a state parameter index, defining the uncertain state output through the information corresponding to the input 3, and judging whether the uncertain state output and the certain state output have interference. Redefining the uncertain state output if the interference exists; and according to the uncertain state output, determining the state output to define various subsequent output information and prompt information.
Step 12, determine if input 3 contains only 0.
If the result is yes, this combination is defined as E, i.e. input 1 is not 0, input 2 is 0, input 3 is 0, state parameter index is defined, and the corresponding deterministic state output is defined by input 1. Outputting and defining various subsequent output information according to the determined state, and defining prompt information;
if the result is negative, an erroneous input, the information is not further processed.
Step 13, determine if input 3 contains 0.
If the result is negative, the combination is defined as H, i.e. input 1 is not 0, input 2 is not 0, input 3 is not 0, state parameter index is defined, indeterminate state output is defined by information corresponding to input 2, input 3, and indeterminate state output is redefined by merging. And judging whether the uncertain state output and the certain state output have interference. Redefining the uncertain state output if the interference exists; and according to the uncertain state output, determining the state output to define various subsequent output information and prompt information.
Step 14, determine if input 3 contains only 0.
If the result is positive, the combination is defined as G, namely input 1 is not 0, input 2 is not 0, input 3 is 0, state parameter index is defined, uncertain state output is defined through information corresponding to input 2, and interference between uncertain state output and certain state output is judged; redefining the uncertain state output if the interference exists; according to the uncertain state output, determining the state output to define various subsequent output information and prompt information;
if the result is no, the information is not further processed for an erroneous input.
In the above steps, as long as the decision classification tree is classified by using a zero set, a non-zero set and an error input as nodes, no matter in which order, different combinations A, B, C, D, … … of three inputs are obtained finally, so that different processing modes can be adopted based on different combinations.
In the above step, it is further preferable that the state parameter index is defined based on:
if all inputs, e.g., input 1, input 2, input 3 in the above example, are 0, the status parameter index is 0;
if all inputs, such as input 1, input 2, input 3 in the above example, are not 0 at the same time, and input 1 is 0, the status parameter index is 1;
if all inputs, e.g. input 1, input 2, input 3 in the above example, are not 0 at the same time, and input 1 is not 0, the status parameter index is 2.
Therefore, all possible processing paths are established through the double setting of the classification decision tree and the state parameter index, and the possibility of interference of deterministic information in the denoising process is fully considered.
Fig. 3 shows a flowchart for classifying and sorting output information of non-zero and uncertain states in an embodiment of the present invention, where the classifying and sorting includes the following steps:
Step 2, because the mechanism is different between different inputs and outputs, the possibility exists that the inputs in the process of deducing the uncertain state output information participate in frequency calculation and do not participate in frequency calculation, so whether each input participates in the frequency calculation related to the sequencing or not needs to be judged here.
And if the input of the defined effective uncertain state output information participates in frequency statistics, judging whether the input is in a negative state, and entering the step 4.
If the definition of the valid uncertain state output information relates to input information which does not participate in frequency statistics, the judgment result is yes, and step 3 is entered.
Step 3, classifying the uncertain state output information based on the correlation among the uncertain state output information defined by different input information, eliminating the uncertain state output information of interference ordering, and defining the uncertain state output information which has no interference and contributes to noise formation;
and 4, calculating the frequency corresponding to each uncertain state output information, and judging whether the uncertain state output information which is free of interference and contributes to noise formation can be sequenced or not.
If the valid uncertain state output information only contains one element, or contains a plurality of elements, and the corresponding frequency of each element is the same, the judgment result is no, no further data processing is needed, the relevant information is output, and the process is ended.
If the effective uncertain state output information contains more than one element and the corresponding frequencies of different uncertain state output information are different, the judgment result is yes;
and 5, sorting the uncertain state output information which does not interfere with the noise formation and contributes to the noise formation according to the frequency, and outputting related information.
Wherein, step 2 is determined by the mechanism of the selected input itself. The expression can be directly simplified according to the logical relationship between the self characteristics of the support data and the support data, and a built-in parameter can be defined to determine the path.
FIG. 4 is a flowchart illustrating an embodiment of output subset processing for non-zero, indeterminate state output information, the processing comprising the steps of:
And 2, judging whether the uncertain state input contributing to noise formation is not zero or not.
If not, terminating the further processing.
And if so, performing data matching on the basis of the related database.
And 3, judging whether the output subsets of the corresponding uncertain states can be sequenced or not.
If not, the step 5 is entered.
And 4, sequencing the corresponding uncertain state output. The sorting step is similar to fig. 3.
And 5, defining output subset related information. The flow is ended.
Fig. 5 is a schematic diagram of the structure of the insert in an embodiment of the invention.
As shown in fig. 5, when the system reads in non-zero uncertain state output information, the system performs three processes at the same time: (1) searching a related database which can be directly processed, and defining subsequent output step by step through data matching; (2) entering a classification and sorting processing step, as shown in fig. 3; (3) the output subset processing step is entered as shown in fig. 4. All of the three processes may be performed, as described in this embodiment, or at least one of the processes may be selectively performed as needed.
Fig. 6 is a schematic diagram illustrating a health information processing flow according to an embodiment of the present invention. The steps of removing interference, defining non-interfering uncertain state output information contributing to noise formation, and sorting are shown in fig. 3. The steps of defining the output subsets and sorting are shown in fig. 4. A plug-in structure requiring parallel processing is shown in fig. 5. This figure shows primarily the overall process, particularly the layer-by-layer peel interference process for noise removal.
The information processing flow comprises the following steps:
And 2, classifying step by step based on the input information. As shown in fig. 2.
And 3, performing data matching, and defining a state parameter index, uncertain state output information, a feedback prompt and other related output information.
And 4, judging whether the state parameter index is not 0.
If the judgment result is negative, the input error is indicated or the input information can not generate effective feedback, and the system outputs the feedback information.
And 5, judging whether the state parameter is not 1.
And if the judgment result is yes, the step 8 is entered.
And 6, judging whether the output information of the uncertain state is not zero.
If the judgment result is no, the existing input information can not provide further information. The corresponding feedback information can be directly output according to the existing information.
If the judgment result is yes, the existing input information can generate a certain amount of information, and the existing information needs further processing.
And 7, forming a parallel processing structure. By inputting different influences of the information on the sorting process, at least one of the following steps is performed simultaneously or selectively: data matching, defining subsequent output step by step (corresponding to the leftmost box in the lower row of fig. 5); classifying the uncertain state output information, and sorting the uncertain state output information (as shown in fig. 3, corresponding to the middle frame in the lower row of fig. 5); the output subsets are defined and ordered by data matching based on the relevant database (as shown in fig. 4, corresponding to the rightmost box in the lower row of fig. 5). And outputting the feedback information.
Step 8, judging whether the output information of the uncertain state is not zero,
if the judgment result is yes, the step 10 is entered.
And 9, performing further data matching through the existing information, defining subsequent output step by step, and outputting all feedback information.
Step 10, judging whether the determined state output interferes with the definition of the uncertain state output information.
If not, go to step 13.
If yes, the process proceeds to step 11 to eliminate the interference information.
And 11, classifying based on the mutual relation between the determined state output with the competitive relation and the initial uncertain state output information, eliminating interference, and defining the relevant information of the effective uncertain state output information.
And step 12, judging whether the output information of the uncertain state is not zero.
If the judgment result is negative, step 14 is entered.
Step 13, a parallel processing structure. The description is the same as for step 7.
And step 14, performing further data matching through the existing information, defining subsequent output step by step, and outputting all feedback information.
The process shown in fig. 6 is to further process the output result preliminarily defined in fig. 2, incorporate into the special situation caused by different characteristics of the data itself, and utilize the steps of fig. 3 to classify and sort step by step, so as to reduce the interference of nonspecific input on the output result to a certain extent, and make the output information more reasonable. The system identification capability is enhanced by the steps and the database arrangement of fig. 4. The information processing capability is enhanced by the structure of fig. 5.
Fig. 6 shows a main flow of health information processing. The specific expression can be finely adjusted according to the characteristics of the research object and the convenient operation. In addition, fig. 6 shows that the system can perform data matching as long as there is a valid primary output, defining the subsequent output stage by stage.
Furthermore, the invention can take the flow as the basis, introduce the time concept, and obtain the information and the sequencing information of the primary output and the primary output subclasses at different time points by accumulating the input. The system obtains various charts or animations based on the information, thereby obtaining dynamic output information and improving the expression and identification capability of the system.
The technical scheme of the invention can be used in the field of medical health, but also can be used in other technical fields containing complex information processing, for example, the scheme of the invention can be used for processing purchasing information for solving complex relationships, such as knowing what goods are sold in nearby shops, supermarkets and malls respectively and the characteristics of the goods, and knowing that some shopping places also provide other special services such as copying and photographing, the complex information processing framework of the application can be used for solving when a user designs a shopping route, and a series of preferred results can be obtained by considering the (certainty) that the user must buy or must deal with, the characteristics of the shops, the availability of goods and the like.
Therefore, the invention constructs an open and extensible framework for efficiently processing complex information. The connecting frame is expandable, so that the application range is expandable, and the problem of limited information flow caused by limited memory and energy of people is solved. The invention improves the processing capability of complex information by comprehensively applying the methods of input setting, classification decision tree, knowledge base processing, noise removal, dynamic output and the like for the first time, shortens the system processing time of the complex information and efficiently completes the generation of personalized output information.
As a preferred embodiment of the present invention, when the present invention is applied to the field of medical health, for example, as a health information expert system, it includes:
a memory storing a program for executing the health information processing method as described above;
a processor for executing programs stored in the memory;
the health information expert system comprises a health information expert system, a health information database and a health information management system, wherein the input information and the output information are information related to the health of a user, the health information expert system further comprises a series of health information databases, and the health information databases store various corresponding relations between the input information and the output information.
The processor may be, for example, a processor in a single chip microcomputer, a desktop, a tablet computer, a notebook computer, a computer network, or a programmable logic controller, as long as the above processing program of the present invention can be executed.
The program can be written according to various common computer languages, such as VB, C + +, JAVA, C #, SQL, MATLAB, R, and so on, as long as the program supports simple database calls and data processing operations.
The health information database, namely the health information knowledge base, can be established in the following way: firstly, structuring the related knowledge of a target system; secondly, generating a series of files according to the corresponding relation between various input information and output information; establishing a relation between input and primary output by adopting a mode of storing information step by step, and storing a corresponding relation between the input and the primary output; establishing a relation between the primary output and the secondary output, and storing the relation between the primary output and the secondary output; and so on until all the target information is reached; all information can be integrated into one file; for systems where the original primary output information is too scattered, such as a health information system, the knowledge base is built by using dimension reduction and expansion processes to improve the recognition capability of the system, i.e., classifying classifiable information, retaining the original name for unclassifiable information, and defining the primary output by using the name of the category, for example, for the output information: park A, park B, park C, park E, park F, park G and park H, wherein park A, park B, park C and park D are all five-star scenic spots, park E, park F and park G are all four-star scenic spots, park H is a three-star scenic spot, and the star-level scenic spot category is taken as primary output; and defining the content contained in each category as a primary output subset, wherein the park A, the park B, the park C and the park D are the primary output subsets aiming at the five-star scenic spot. Through the processing, the system can process corresponding information from a global layer and a local layer, improve the identification capability of the system and optimize an output result.
The health information database can only store input information keywords, output information keywords and corresponding relations between the input information and the output information, can also store data groups of a plurality of items, and searches the input information keywords in different items of the data groups, so that more accurate searching and matching can be realized, and the input/output information can be respectively placed in different databases and matched and called through corresponding relation sets.
As a specific example, the health information expert system of the present invention can systematize and personalize information, so that a user can efficiently and economically obtain relevant information adapted to his/her needs.
For example, for the health information field, if some assay indicator has already made it clear what disease, then certain status input information may be obtained, defined as the first input; the possibility of the disease is greatly increased although it is not certain whether the disease will be caused, based on family genetic history, etc., and it can be defined as the second input; the common symptoms are defined as the third input, since each symptom may correspond to many aspects of the disease. Thus, when the user has a cardiovascular problem (example 1) and wishes to know the relevant information, the heart disease can be entered in the first input, and the system will prompt the user to pay attention to blood pressure, blood sugar, blood fat, and the relevant information; if the user wants to learn information about shortness of breath and bluing fingernails, the keyword set can be entered in the third input and the system will prompt a preference for lung problems (example 2).
The system output results are affected by the input quality. If the input information is not comprehensive and accurate, the accuracy of the output result is poor. In addition, the current denoising method is to partially denoise the noise under the condition of limited information quantity so as to improve the output accuracy. Taking example 2 as an example, the system prompts the user to pay attention to lung lesions by sorting from a series of possible lesions, so that the personalized information obtained by the user is more accurate. In addition, the system suggests a highly likely pathology to become a chronic obstructive lung.
Example 1
In the first input "heart disease" is entered.
This example lists some of the run results for a heart user. Based on the user's output, the system lists the lesion as a cardiac lesion. The system prompts the user to pay attention to control the problems of hypertension, hyperglycemia, hyperlipidemia and the like according to the extracted information. In addition, the system prompts the user to stop smoking or avoid second-hand smoking, maintain a healthy diet, pay attention to exercise, and the like.
The identified diseases needed to work with Dr.：
1.heart disease
keep levels of the following items under control：
1.high blood pressure
2.diabetes
3.high blood cholester。l/triglycerides
…
Pay attention to the following issues：
1.quit smoking and avoid secondhand smoking
2.keep healthy diet
3.keep regular exercise
The above output results are only schematic illustrations of the output of the present application, and the language and the text used therein can be adjusted according to different applications.
Example 2
In the third input, "qibrachi and fingernail bluing" is input.
In this example, the user is listed with the results of some operations by taking qi shortness and bluing fingernails as examples. Based on the user input, the system lists the possible lesions as heart lesions, kidney lesions, lung lesions, stomach lesions, etc., with the highest likelihood of lung lesions. The system prompts the user to pay attention to control the problems of hypertension, hyperglycemia, hyperlipidemia and the like according to the extracted information. In addition, the system prompts the user to stop smoking or avoid second-hand smoke, limit alcohol consumption, maintain healthy diet and other things. By outputting the subclasses, the system can also list the highest likelihood of a slow obstructive lung among lung lesions. Since this run lists all possibilities, the user can re-run the system based on the sorted results with lung lesions as input to obtain relatively clear, legible results.
Pay more attention to(if possible，go to check)the following diseases：
1.heart disease
2.kidney disease
3.lung disease
4.stomach disease
5.lung cancer
Potential diseases with probability from high to low：
lung disease＞heart disease，kidney disease，stomach disease，lung cancer
keep levels of the following items under control：
1.high blood pressure
2.diabetes
3.high blood cholesterol/triglycerides
Pay attention to the following issues：
1.quit smoking and avoid secondhand smoking
2.limit alcohol
3.keep healthy diet
Applications of the invention
As can be seen from the above description, the technology of the present invention can be used in many technical fields requiring the processing of miscellaneous information, such as the field of medical health, and particularly the field of health information expert systems. In real life, health related information is extremely rich and scattered, and the general public has great demands on health information. The health information is highly complex, and the existing health information supply mode is low in efficiency, so that the high-quality health information is relatively insufficient. The medical expert system can systematize and personalize the information, so that a user can efficiently and economically obtain the relevant information meeting the self requirement. Therefore, the health information processing system and the health information expert system are important supplements of the health system, the problems of high cost and low efficiency of the current health information supply are solved through technical means, the popularization of the health information is realized, and the maintenance of the health of ordinary people is of important value. In addition, the health information processing system and the health information expert system can provide a relatively systematic visual angle for professionals with relatively insufficient knowledge and experience, and the occurrence of misjudgment is reduced. As the system becomes more sophisticated, other advantages will become apparent.
It should be noted that the output data and the corresponding relationship between various inputs and symptoms in the present application are derived from the known sources of those skilled in the art, because health science is a science, and the data has objective certainty. In addition, the present application is not a disease diagnosis and treatment system, which provides various possibilities for ranking according to the known correspondence between the external manifestations of symptoms and diseases, and further provides expanded knowledge of the cautionary matters and general nursing knowledge of diseases, and is more inclined to provide auxiliary reference information.
The above-mentioned embodiments are intended to illustrate the objects, technical solutions and advantages of the present invention in further detail, and it should be understood that the above-mentioned embodiments are only exemplary embodiments of the present invention and are not intended to limit the present invention, and any modifications, equivalents, improvements and the like made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (8)
1. A health information processing method, characterized by comprising the steps of:
step 1, performing decision tree classification on input information, wherein the input information is classified based on two aspects of whether a deterministic relationship exists between the input information and output information and whether the input information in an uncertain state contributes to noise, and the input information is in a set form;
step 2, matching the input information of different categories with corresponding output information according to different input combinations obtained by the decision tree classification in the step 1, and storing the input information as an intermediate result; wherein the output information can also be in the form of a set;
step 3, selecting a corresponding processing mode for the intermediate result according to different input combinations obtained by the decision tree classification in the step 1 to obtain corresponding final output information;
the health information expert system comprises a health information expert system, a health information database and a health information management system, wherein the input information and the output information are both information related to health, the health information expert system further comprises a plurality of health information databases, and various input and output corresponding relations are stored in the health information databases;
the health information database is capable of defining primary outputs and primary output subsets on demand;
the decision tree classification step in step 1 specifically includes:
defining input information with certainty as first input information, wherein output information obtained by matching the first input information is first output information;
defining input information which has no determinacy and does not contribute to noise formation as second input information, wherein output information obtained by matching the second input information is second output information;
defining input information which has no determinacy and contributes to noise formation as third input information, wherein output information obtained by matching the third input information is third output information; the second output information and the third output information are jointly called uncertain state output information;
carrying out decision tree classification by taking whether the first input information, the second input information and the third input information are zero sets, non-zero sets and error inputs as nodes;
wherein, a null set is defined as a set that contains only one element and is 0, indicating no information amount, a non-null set is defined as a set that contains at least one element and all elements are not 0, indicating valid information, and if a set contains both 0 and other elements, this set is defined as an erroneous input.
2. The health information processing method according to claim 1,
the different input combinations obtained by the decision tree classification in the step 1 are specifically obtained based on the following rules:
when all the input information is 0, setting a state parameter index and assigning the state parameter index to be 0;
when all the input information are not 0 at the same time and the first input information comprises 0, setting a state parameter index and assigning a value to be 1;
when all input information is not 0 at the same time and the first input information does not include 0, setting a state parameter index and assigning a value of 2;
thus, the step 3 specifically includes:
step 301, judging whether the state parameter index is not 0;
if the judgment result is negative, the input error is indicated or the input information can not generate effective feedback, the system outputs feedback information, and the whole method flow is ended;
step 302, judging whether the state parameter index is not 1;
if the judgment result is yes, go to step 305;
step 303, judging whether the output information of the uncertain state is not null; if the judgment result is negative, the existing input information can not provide further information, and the whole method flow is ended;
and step 304, simultaneously or selectively executing at least one of the following steps through different influences of the input information on the sorting process: firstly, data matching, defining subsequent output step by step, classifying and sequencing the output information of the uncertain state, and thirdly, defining and sequencing an output subset through data matching on the basis of a relevant database; after the execution of the first step and the third step is finished, feedback information is output, and the whole method flow is finished;
step 305, judging whether the output information of the uncertain state is not null;
if the judgment result is negative, defining subsequent output step by step through data matching; outputting a feedback result and finishing the whole method flow;
step 306, judging whether the first output information has an inclusion or repetition relationship with the content in the uncertain state output information as an intermediate result;
if not, go to step 309;
if yes, deleting the content with the inclusion or repetition relationship in the uncertain state output information;
step 307, classifying the contents with mutual competition relationship in the second output information and the third output information, and further deleting the contents with mutual competition relationship and repetition from the uncertain state output information of the deleted contents with or repetition relationship;
step 308, judging whether the output information of the uncertain state obtained by the step 307 is not null;
if the judgment result is negative, defining subsequent output step by step through data matching; outputting a feedback result and finishing the whole method flow;
step 309, by inputting different influences of the information on the sorting process, at least one of the following steps is performed simultaneously or selectively: firstly, data matching, defining subsequent output step by step, classifying and sequencing the output information of the uncertain state, and thirdly, defining and sequencing an output subset through data matching on the basis of a relevant database; after the execution is finished, outputting feedback information, and finishing the whole process;
step 310, outputting all feedback information, and ending the whole method flow;
in steps 304 and 309, the step of performing output subset processing on the non-zero uncertain state output information specifically includes:
step 3041, the system reads in non-zero uncertain state output information;
step 3042, determining whether the uncertain state input information contributing to noise formation is not zero;
if the judgment is no, the further processing is terminated;
if yes, carrying out data matching on the basis of the related database;
step 3043, determining whether the output subsets corresponding to the uncertain states can be sorted;
if not, go to step 3045;
step 3044, sorting the corresponding uncertain state outputs;
step 3045, defining the relevant information of the output subset, and ending the whole process.
3. The health information processing method according to claim 2, wherein when the judgment result is an input error or the input information cannot generate a valid feedback, the existing information or the corresponding feedback information set in advance is directly output.
4. The health information processing method according to claim 3, wherein the step 310 specifically comprises the steps of:
step 3101, according to the difference of the characteristics of the second output information, determining whether each second output information contributes to noise formation and participates in the frequency calculation related to the ranking;
if the second output information and the third output information both contribute to noise formation and the results can be combined, judging whether the results are negative; proceed to step 3103;
step 3102, classifying based on the relationship between the uncertain state output information matched with different input information, eliminating the uncertain state output information of interference ordering, and obtaining redefined uncertain state output information;
step 3103, calculating the frequency corresponding to each uncertain state output information, and determining whether the non-interfering uncertain state output information can be sorted;
if the effective output information of the uncertain state only contains one element, or contains a plurality of elements but the corresponding frequency of each element is the same, judging whether the result is negative, outputting the relevant information without further data processing, and ending the whole process;
and 3104, sorting the interference-free uncertain state output information according to the frequency, and outputting the related information.
5. The health information processing method according to claim 4, wherein step 3101 is determined by the characteristics of the selected input; the expression is directly simplified according to the logical relation between the self characteristics of the support data and the support data, or a built-in parameter is defined to determine the path.
6. The health information processing method according to claim 1, wherein the input information is selected from preset options, or is input through a keyboard or voice;
selecting the preset option through a drop-down list or through single-selection or multi-selection preset grouping options;
grouping preset options according to the inherent rule of the preset options, and supporting the association function of the options;
in addition, the system can input a series of input information based on a time axis and output comprehensive output information based on the time axis through circularly executing the steps 1-3;
the health information processing method further comprises the following step 4: in addition, inputting a series of input information based on a time axis, and circularly executing the steps 1-3 to obtain comprehensive output information based on the time axis, wherein the original output information of the steps 1-3 is shielded; the health information processing method further includes a step 5 of outputting the obtained integrated output information in the form of a graph or animation as dynamic output information related to a time axis.
7. A health information processing system, comprising:
a memory storing a program for executing the health information processing method according to any one of claims 1 to 6;
a processor for executing programs stored in the memory;
the health information processing system also comprises a plurality of information databases, and the information databases store various input and output corresponding relations;
the information database can define primary output and primary output subsets according to requirements in the construction process.
8. The health information processing system of claim 7, wherein the processor is a processor in a single chip, a desktop, a tablet, a laptop, a computer network, or a programmable logic controller.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810232176.8A CN110310739B (en) | 2018-03-20 | 2018-03-20 | Health information processing method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810232176.8A CN110310739B (en) | 2018-03-20 | 2018-03-20 | Health information processing method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110310739A CN110310739A (en) | 2019-10-08 |
| CN110310739B true CN110310739B (en) | 2022-06-24 |
Family
ID=68073605
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810232176.8A Active CN110310739B (en) | 2018-03-20 | 2018-03-20 | Health information processing method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110310739B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101261637A (en) * | 2007-03-08 | 2008-09-10 | 国际商业机器公司 | System and method of searching key word for determining event happening |
| CN105912645A (en) * | 2016-04-08 | 2016-08-31 | 上海智臻智能网络科技股份有限公司 | Intelligent question and answer method and apparatus |
| WO2017057921A1 (en) * | 2015-10-02 | 2017-04-06 | 네이버 주식회사 | Method and system for automatically classifying data expressed by a plurality of factors with values of text word and symbol sequence by using deep learning |
| CN107403067A (en) * | 2017-07-31 | 2017-11-28 | 京东方科技集团股份有限公司 | Intelligence based on medical knowledge base point examines server, terminal and system |
| CN107729319A (en) * | 2017-10-18 | 2018-02-23 | 百度在线网络技术（北京）有限公司 | Method and apparatus for output information |
| CN107784057A (en) * | 2017-03-03 | 2018-03-09 | 平安医疗健康管理股份有限公司 | Medical data matching process and device |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| MXPA02006733A (en) * | 2000-01-06 | 2004-09-10 | Igotpain Com Inc | System and method of decision making. |
| CN101127100A (en) * | 2006-08-18 | 2008-02-20 | 张湛 | Construction method for intelligent system for processing uncertain cause and effect relationship information |
| CN101751509A (en) * | 2008-12-12 | 2010-06-23 | 北京大北农科技集团股份有限公司 | Swine disease diagnosis system and swine disease diagnosis method |
| CN103955608B (en) * | 2014-04-24 | 2017-02-01 | 上海星华生物医药科技有限公司 | Intelligent medical information remote processing system and processing method |
| US9390706B2 (en) * | 2014-06-19 | 2016-07-12 | Mattersight Corporation | Personality-based intelligent personal assistant system and methods |
| CN107590166B (en) * | 2016-12-20 | 2019-02-12 | 百度在线网络技术（北京）有限公司 | A kind of data creation method and device based on inquiry content |
-
2018
- 2018-03-20 CN CN201810232176.8A patent/CN110310739B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101261637A (en) * | 2007-03-08 | 2008-09-10 | 国际商业机器公司 | System and method of searching key word for determining event happening |
| WO2017057921A1 (en) * | 2015-10-02 | 2017-04-06 | 네이버 주식회사 | Method and system for automatically classifying data expressed by a plurality of factors with values of text word and symbol sequence by using deep learning |
| CN105912645A (en) * | 2016-04-08 | 2016-08-31 | 上海智臻智能网络科技股份有限公司 | Intelligent question and answer method and apparatus |
| CN107784057A (en) * | 2017-03-03 | 2018-03-09 | 平安医疗健康管理股份有限公司 | Medical data matching process and device |
| CN107403067A (en) * | 2017-07-31 | 2017-11-28 | 京东方科技集团股份有限公司 | Intelligence based on medical knowledge base point examines server, terminal and system |
| CN107729319A (en) * | 2017-10-18 | 2018-02-23 | 百度在线网络技术（北京）有限公司 | Method and apparatus for output information |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110310739A (en) | 2019-10-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107609101B (en) | Intelligent interaction method, equipment and storage medium | |
| KR101778679B1 (en) | Method and system for classifying data consisting of multiple attribues represented by sequences of text words or symbols using deep learning | |
| CN116756579B (en) | Training method of large language model and text processing method based on large language model | |
| US20060173753A1 (en) | Method and system for online shopping | |
| US8156138B2 (en) | System and method for providing targeted content | |
| EP2757510A1 (en) | Method and system for linking data sources for processing composite concepts | |
| US20200388358A1 (en) | Machine Learning Method for Generating Labels for Fuzzy Outcomes | |
| CN103455538B (en) | Information processing unit, information processing method and program | |
| CN113674864B (en) | Malignant tumor combined venous thromboembolism risk prediction method | |
| TW201232301A (en) | Identifying categorized misplacement | |
| CN111353851A (en) | Hotel sorting recommendation method and device, electronic equipment and storage medium | |
| US11734589B2 (en) | Virtual assistant negotiation system and method thereof | |
| WO2013103747A1 (en) | Detecting overlapping clusters | |
| US20220383162A1 (en) | Method and system for data classification to generate a second alimentary provider | |
| US20230222364A1 (en) | Method of and system for generating a longevity element and an instruction set for a longevity element plan | |
| TW202133080A (en) | Resource information pushing method and device, server and storage medium | |
| CN103186650B (en) | A kind of searching method and device | |
| US20220138619A1 (en) | Method and system for generating an alimentary element prediction machine-learning model | |
| CN110310739B (en) | Health information processing method and system | |
| Rossetti et al. | Forecasting success via early adoptions analysis: A data-driven study | |
| CN112861956A (en) | Water pollution model construction method based on data analysis | |
| CN113032676A (en) | Recommendation method and system based on micro-feedback | |
| CN107807949A (en) | Intelligent interactive method, equipment and storage medium | |
| CN109885668A (en) | A kind of expansible field interactive system status tracking method and apparatus | |
| CN115719640A (en) | System, device, electronic equipment and storage medium for recognizing primary and secondary symptoms of traditional Chinese medicine |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP02 | Change in the address of a patent holder | ||
| CP02 | Change in the address of a patent holder |
Address after: Room 503, Unit 2, Building 7, No. 350 Fanyang Middle Road, Zhuozhou City, Baoding City, Hebei Province Patentee after: He Lijun Address before: Room 602, unit 1, building 2, Yiming garden, Chengnan Jiayuan, Fengtai District, Beijing 100068 Patentee before: He Lijun |