Automatic medical consultation method and device

Abstract

The present invention discloses an automatic medical inquiry method and device, the method comprising: receiving user input information; extracting information from the input information to obtain symptom information; determining whether the symptom information is complete; if not, inquiring about the missing information in the symptom information, and completing the information according to the user's answer until the complete symptom information is obtained; determining the predicted disease according to the complete symptom information; if there is only one predicted disease, obtaining and outputting the medical inquiry result; if there are multiple predicted diseases, obtaining the inquiry information corresponding to the predicted disease, and outputting the inquiry information. The present invention can improve the efficiency of medical inquiry and the accuracy of medical inquiry results.

Description

Technical Field

[0001] The present invention relates to the field of question-answering systems, and in particular to an automatic diagnosis method and device. Background Art

[0002] The question-answering system is an advanced form of information retrieval system. It can not only provide users with a more natural human-computer interaction method, but also directly return the answer to the user's question, helping users to quickly and accurately obtain the required information. It is currently a research direction in the field of artificial intelligence and natural language processing that has attracted much attention and has broad development prospects. Different applications require different forms of question-answering systems. For this reason, some automatic consultation systems have emerged in the medical field. However, most of the existing automatic consultation products are based on the information input by the user, giving options for the user to choose, and then giving new options based on the user's choice until the consultation result is given at the end. This method has the following disadvantages: 1) The process is lengthy and boring, and the efficiency is low; 2) The options given sometimes cannot accurately describe the user's problem, resulting in low accuracy of the consultation results. Summary of the invention

[0003] The embodiment of the present invention provides an automatic medical consultation method and device to improve the medical consultation efficiency and the accuracy of medical consultation results.

[0004] To this end, the present invention provides the following technical solutions:

[0005] An automatic medical consultation method, comprising:

[0006] Receive user input information;

[0007] Extracting information from the input information to obtain symptom information;

[0008] Determining whether the disease information is complete;

[0009] If not, inquire about the missing information in the symptom information, and complete the information according to the user's answer until the complete symptom information is obtained;

[0010] Determining the predicted disease based on the complete symptom information;

[0011] If there is only one predicted disease, obtaining and outputting the consultation result;

[0012] If there are multiple predicted diseases, query information corresponding to the predicted diseases is obtained and the query information is output.

[0013] Optionally, determining whether the disease information is complete includes:

[0014] Determining whether a unique disease type can be determined based on the symptom information;

[0015] If yes, it is determined that the disease information is complete;

[0016] Otherwise, it is determined that the disease information is incomplete.

[0017] Optionally, the method further comprises:

[0018] Obtaining all symptoms of the candidate disease determined according to the symptom information;

[0019] Determine the discrimination degree and weight of each symptom, and calculate the score of the symptom according to the discrimination degree and weight of the symptom;

[0020] The querying of missing information in the symptom information includes:

[0021] The highest scoring symptoms were selected for questioning.

[0022] Optionally, the method further comprises:

[0023] If no symptom information is obtained by extracting the input information, the user is guided to re-enter the information.

[0024] Optionally, guiding the user to re-input includes:

[0025] Prompt the user to re-enter details; or

[0026] Ask the user a question to make the user re-enter the information based on the question.

[0027] Optionally, determining the predicted disease according to the complete symptom information includes:

[0028] According to the complete symptom information, a pre-established medical knowledge graph is searched to determine the predicted disease.

[0029] Optionally, there are multiple predicted diseases; the method further includes:

[0030] A predicted disease is selected from the multiple predicted diseases using a pre-established entity prediction model.

[0031] Optionally, the obtaining of inquiry information corresponding to the predicted disease includes:

[0032] All query information of the plurality of predicted diseases is obtained from the medical knowledge graph, and non-common query information corresponding to each predicted disease is determined.

[0033] Optionally, the outputting the query information includes: simultaneously outputting non-common query information corresponding to each predicted disease;

[0034] The method further comprises:

[0035] After receiving the user's input information for the non-common query information, extracting the input information to obtain auxiliary information;

[0036] Determining whether the auxiliary information is complete;

[0037] If yes, determining and outputting the consultation result according to the complete auxiliary information;

[0038] Otherwise, prompt the user to re-enter.

[0039] Optionally, the outputting the inquiry information includes:

[0040] Determining the probability of each predicted disease among the plurality of predicted diseases;

[0041] Output non-common query information corresponding to a predicted disease in order from high to low probability;

[0042] The method further comprises:

[0043] After receiving the user's input information for the non-common information, extracting the input information to obtain supplementary information;

[0044] Determining whether the supplemental information is complete;

[0045] If the supplementary information is incomplete, continue to output the next non-common inquiry information of the predicted disease until all the non-common inquiry information is output;

[0046] After the complete supplementary information is obtained, the medical consultation result is determined and outputted according to the complete supplementary information.

[0047] Optionally, the outputting of the consultation result includes:

[0048] Converting the medical consultation results into natural language text;

[0049] The natural language text is output.

[0050] Optionally, converting the medical consultation result into natural language text includes:

[0051] Filling the medical consultation result into a preset template to obtain a natural language text; or

[0052] The inquiry result is input into a pre-established inquiry model, and a natural language text is obtained according to the output of the inquiry model.

[0053] An automatic medical consultation device, comprising:

[0054] A receiving module, used for receiving user input information;

[0055] An information extraction module, used to extract information from the input information to obtain symptom information;

[0056] A judgment module, used to determine whether the disease information is complete;

[0057] A feedback module, used for, when the judgment module determines that the symptom information is incomplete, inquiring about the missing information in the symptom information, and completing the information according to the user's answer until the complete symptom information is obtained;

[0058] A prediction module, used to determine the predicted disease based on the complete symptom information;

[0059] An information acquisition module, used to obtain the consultation result when there is only one predicted disease; and to obtain the inquiry information corresponding to the predicted disease when there are multiple predicted diseases;

[0060] An output module is used to output the medical inquiry result or the inquiry information.

[0061] Optionally, the judgment module is specifically used to determine whether a unique disease type can be determined based on the symptom information; if so, determine that the symptom information is complete; otherwise, determine that the symptom information is incomplete.

[0062] Optionally, the device further comprises:

[0063] A symptom acquisition module, used to acquire all symptoms of the candidate disease determined according to the symptom information;

[0064] A score determination module, used to determine the discrimination and weight of each symptom, and calculate the score of the symptom according to the discrimination and weight of the symptom;

[0065] The feedback module selects the symptoms with the highest scores for inquiry.

[0066] Optionally, the device further comprises:

[0067] The guiding module is used to guide the user to re-enter the information when the information extraction module fails to obtain the symptom information by extracting the input information.

[0068] Optionally, the guiding module is specifically used to guide the user to re-enter in the following manners: prompting the user to re-enter detailed content; or asking questions to the user so that the user re-enters according to the question content.

[0069] Optionally, the prediction module includes: a graph query module, which is used to search a pre-established medical knowledge graph based on the complete symptom information to determine the predicted disease.

[0070] Optionally, there are multiple predicted diseases; the prediction module also includes:

[0071] The filtering module is used to select a predicted disease from the multiple predicted diseases using a pre-established entity prediction model.

[0072] Optionally, the information acquisition module is specifically used to obtain all query information of the multiple predicted diseases from the medical knowledge graph, and determine non-common query information corresponding to each predicted disease.

[0073] Optionally, the output module is further used to simultaneously output non-common query information corresponding to each predicted disease;

[0074] The information extraction module is further configured to extract information from the input information of the user for the non-common inquiry information after the receiving module receives the input information of the user to obtain auxiliary information;

[0075] The judging module is further used to determine whether the auxiliary information is complete;

[0076] The prediction module is further configured to determine the final disease according to the complete auxiliary information when the judgment module determines that the auxiliary information is complete;

[0077] The information acquisition module is further used to obtain the diagnosis result corresponding to the final disease;

[0078] The output module is further used to output the consultation result;

[0079] The guiding module is further used to guide the user to re-enter the auxiliary information when the judging module determines that the auxiliary information is incomplete.

[0080] Optionally, the device further comprises:

[0081] A probability calculation module, used to determine the probability of each predicted disease among the multiple predicted diseases;

[0082] The output module is also used to output non-common query information corresponding to a predicted disease in order from high to low probability;

[0083] The information extraction module is further configured to extract information from the input information of the user for the non-common information after the receiving module receives the input information of the user for the non-common information to obtain supplementary information;

[0084] The judging module is further used to determine whether the supplementary information is complete;

[0085] The output module is further configured to continue outputting the next non-common inquiry information of the predicted disease when the judgment module determines that the supplementary information is incomplete, until all the non-common inquiry information is outputted;

[0086] The prediction module is further used to determine the final disease according to the complete supplementary information after obtaining the complete supplementary information;

[0087] The information acquisition module is further used to obtain the diagnosis result corresponding to the final disease;

[0088] The output module is also used to output the medical consultation result.

[0089] Optionally, the output module includes:

[0090] A text conversion unit, used for converting the inquiry information or the medical inquiry result into a natural language text;

[0091] A text output unit is used to output the natural language text.

[0092] Optionally, the text conversion unit includes:

[0093] A template filling unit, used to fill the inquiry information or the medical inquiry result into a preset template to obtain a natural language text; or

[0094] The text generation unit is used to input the inquiry information or the medical inquiry result into a pre-established inquiry model, and obtain a natural language text according to the output of the inquiry model.

[0095] A computer device, comprising: one or more processors, a memory;

[0096] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the aforementioned method.

[0097] A readable storage medium stores instructions thereon, wherein the instructions are executed to implement the above-mentioned method.

[0098] The automatic medical consultation method and device provided by the embodiment of the present invention extracts information from the input information of the user to obtain the symptom information after receiving the input information of the user. If the extracted symptom information is incomplete, the missing information is queried and the information is supplemented according to the user's answer until the complete symptom information is obtained. For the complete symptom information, the pre-established medical knowledge graph is used to search for possible predicted diseases, and then the corresponding medical consultation results or inquiry information are obtained according to the predicted disease and output. The medical consultation results are determined and output according to the inquiry information.

[0099] Furthermore, in the case where the predicted diseases include multiple diseases, non-common inquiry information of these predicted diseases is displayed to the user, and corresponding auxiliary information or supplementary information is obtained based on the user's input information for the non-common inquiry information, and then the final inquiry result is obtained and output based on the auxiliary information or supplementary information. BRIEF DESCRIPTION OF THE DRAWINGS

[0100] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0101] Figure 1 is a flow chart of the automatic medical consultation method according to an embodiment of the present invention;

[0102] Figure 2 is another flow chart of the automatic diagnosis method according to an embodiment of the present invention;

[0103] Figure 3 is another flow chart of the automatic diagnosis method according to an embodiment of the present invention;

[0104] Figure 4 is a structural block diagram of an automatic medical consultation device according to an embodiment of the present invention;

[0105] Figure 5 is a block diagram of a device for an automatic medical consultation method according to an exemplary embodiment;

[0106] Figure 6 It is a schematic diagram of the structure of the server in the embodiment of the present invention. DETAILED DESCRIPTION

[0107] In order to enable persons skilled in the art to better understand the solutions of the embodiments of the present invention, the embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings and implementation modes.

[0108] The embodiment of the present invention provides an automatic medical consultation method and system. After receiving the user's input information, the input information is extracted to obtain the symptom information. If the extracted symptom information is incomplete, the missing information is queried, and the information is supplemented according to the user's answer until the complete symptom information is obtained. For the complete symptom information, the pre-established medical knowledge graph is used to find possible predicted diseases, and then the corresponding query information is obtained according to the predicted disease, and the medical consultation result is determined and output according to the query information.

[0109] like Figure 1 FIG. 1 is a flow chart of an automatic diagnosis method according to an embodiment of the present invention, which includes the following steps:

[0110] Step 101: receiving user input information.

[0111] Step 102: extract the input information to obtain symptom information.

[0112] The user's input information is usually in a natural language mode, and some relevant information required for disease diagnosis needs to be extracted from it. For the convenience of description, the relevant information is called symptom information.

[0113] In the embodiment of the present invention, the symptom information includes user information and condition information. The user information includes but is not limited to any one or more of the following information: gender, age, and may further include some auxiliary information that is helpful for judging the disease, such as living environment and type of work; the condition information includes but is not limited to: symptoms, medical history and other information.

[0114] Information extraction can use existing technologies, such as using the CRF++ model to extract relevant medical entities. In addition, considering that sometimes there may be multiple different expressions or corresponding terms for the same disease, in order to make the extracted disease information more standardized, the extracted disease information can be further standardized using relevant dictionaries (such as synonym dictionaries, synonym dictionaries, etc.) and TFIDF+SVM models to obtain standardized disease information.

[0115] For example, the user's input information is "I have a terrible headache", and the extracted disease information is: "headache", marked as "symptoms".

[0116] Step 103 , determining whether the disease information is complete; if so, executing step 104 ; otherwise, executing step 109 .

[0117] The criterion for judging whether the symptom information is complete can be determined according to the characteristics of each disease, and specifically can be judged according to whether the extracted symptom information can determine a unique disease type. The disease types can be divided according to different standards, such as hospital departments, national standards, international standards, etc., which are not limited in the embodiments of the present invention. For example, according to the site of onset, it can be divided into eyes, hands, feet, heart, etc.; according to the department, it can be divided into ophthalmology, internal medicine, respiratory department, etc.; according to the onset group, it can be divided into diseases of different age groups and different genders; according to the cause of the disease, it can be divided into bacterial infection, fungal infection, trauma, etc.; according to the infectiousness, it can be divided into highly infectious diseases, low infectious diseases, etc.

[0118] Step 104, determining the predicted disease based on the complete symptom information.

[0119] In an embodiment of the present invention, a medical knowledge graph may be established in advance, and the disease corresponding to the complete symptom information, that is, the predicted disease, may be determined by searching the medical knowledge graph.

[0120] The medical knowledge graph is a purely knowledge-based graph that can be established through existing medical knowledge and relationships.

[0121] Its content includes but is not limited to the following: disease, symptoms of the disease, gender and age corresponding to the disease, inquiry information about the disease (such as medical history, type of work, living habits, etc.), etc.

[0122] It should be noted that, since some different diseases may have some of the same symptoms, there may be one or more predicted diseases obtained by searching the medical knowledge graph based on the complete symptom information. If there is only one, then only this one disease can be queried later; if there are multiple, a pre-established entity prediction model can be used to select one predicted disease from the multiple predicted diseases. Specifically, the entity prediction model is used to calculate the probability of each disease, and then the one with the highest probability is used as the predicted disease found. Because some different diseases with the same symptoms have different probabilities of occurrence, for example, the probabilities of common diseases and rare diseases can be very different. The entity prediction model can be trained using existing medical question-and-answer data.

[0123] In addition, it should be noted that in actual applications, the user's input information may also contain information about symptoms that the user denies. For this reason, in another embodiment of the method of the present invention, this information can also be extracted as reference information. When searching using the medical knowledge graph, this reference information can be combined to exclude diseases containing symptoms corresponding to the reference information.

[0124] Step 105, determining whether there is only one predicted disease. If yes, proceed to step 106; otherwise, proceed to step 107.

[0125] Step 106, obtaining and outputting the medical consultation result.

[0126] The consultation result can be obtained from the medical knowledge graph.

[0127] Step 107, obtaining query information corresponding to the predicted disease.

[0128] Specifically, the query information corresponding to the predicted disease can be directly obtained from the medical knowledge graph.

[0129] Step 108: output the inquiry information.

[0130] Step 109, inquiring about the missing information in the symptom information; then executing steps 101 to 103, that is, continuing to receive the user's input information, and completing the information according to the user's answer until the complete symptom information is obtained.

[0131] Based on the extracted symptom information, it may be determined that there are multiple possible diseases. In this case, you can find the most distinguishing and common symptoms to ask questions. High distinction helps reduce the number of conversation turns to improve efficiency, and high commonness helps users answer and improve user experience.

[0132] Specifically, all symptoms of the candidate disease determined based on the symptom information may be obtained; the discrimination and weight of each symptom may be determined; the score of the symptom may be calculated based on the discrimination and weight of each symptom; and then the symptom with the highest score may be selected for inquiry.

[0133] The discrimination and weight of each symptom can be determined as follows: Assume that the number of diseases is a. Find the commonality of all symptoms related to the multiple possible diseases, and process these commonality with the Softmax function, and the result is used as the weight w of each symptom. Then count the number of times each symptom appears in all diseases, assuming it is b, then the discrimination corresponding to each symptom is ab, and the discrimination of all symptoms is processed with the Softmax function, and then multiplied by the corresponding weight w to obtain the score of each symptom.

[0134] The discrimination and weight of each symptom can also be determined in the following manner: the distribution of each symptom relative to a variety of possible diseases is calculated using a statistical distribution method, and the distribution is used for calculation to obtain the discrimination and weight of each symptom.

[0135] In some cases, no disease information may be extracted from the user's first round of input information. In this case, the user can be guided to re-enter the information, for example, by prompting the user to re-enter the details; or by asking the user a question so that the user can re-enter the information based on the question content.

[0136] It should be noted that in the above step 104, when the predicted disease is searched and determined according to the medical knowledge graph, there may be one or more predicted diseases found.

[0137] If only one predicted disease is found, then in step 106, the inquiry information corresponding to the predicted disease can be directly output, that is, the inquiry information is used as the final consultation result.

[0138] If there are multiple predicted diseases found, then in the above step 107, it is necessary to obtain all query information of the multiple predicted diseases from the medical knowledge graph.

[0139] Accordingly, when outputting the query information, multiple methods are used. Figure 2 and Figure 3 Provide detailed explanation.

[0140] like Figure 2 FIG. 2 is another flow chart of the automatic diagnosis method according to an embodiment of the present invention, which includes the following steps:

[0141] Step 201: receiving user input information.

[0142] Step 202: extract the input information to obtain symptom information.

[0143] Step 203 , determining whether the disease information is complete; if so, executing step 204 ; otherwise, executing step 207 .

[0144] Step 204, determining predicted diseases based on the complete symptom information to obtain multiple predicted diseases.

[0145] Step 205: Obtain all query information of the multiple predicted diseases, and determine non-common query information corresponding to each predicted disease.

[0146] Step 206, simultaneously outputting non-common query information corresponding to each predicted disease.

[0147] Step 207: after receiving the user's input information for the non-common query information, extract the input information to obtain auxiliary information.

[0148] Step 208, determining whether the auxiliary information is complete. If yes, executing step 209; otherwise, executing step 210.

[0149] It should be noted that, for different diseases, the amount and content of the auxiliary information may be different, and may be specifically determined according to the needs of various diseases.

[0150] Step 209: Determine and output the medical consultation result based on the complete auxiliary information.

[0151] Step 210: Instruct the user to re-enter the information, and then return to step 207.

[0152] Step 211, inquire about the missing information in the symptom information; then execute steps 201 to 203, that is, continue to receive the user's input information, and complete the information according to the user's answer until the complete symptom information is obtained.

[0153] In this embodiment, after searching for multiple predicted diseases, non-common inquiry information of these predicted diseases is output simultaneously. Then, based on the user's input information for the non-common inquiry information, complete auxiliary information can be extracted, and the final disease can be determined using the auxiliary information to obtain the corresponding consultation results.

[0154] like Figure 3 FIG. 2 is another flow chart of the automatic diagnosis method according to an embodiment of the present invention, which includes the following steps:

[0155] Step 301: receiving user input information.

[0156] Step 302: extract the input information to obtain symptom information.

[0157] Step 303 , determining whether the disease information is complete; if so, executing step 304 ; otherwise, executing step 307 .

[0158] Step 304, determining predicted diseases based on the complete symptom information to obtain multiple predicted diseases.

[0159] Step 305: Obtain all query information of the multiple predicted diseases, and determine non-common query information corresponding to each predicted disease.

[0160] Step 306, determining the probability of each predicted disease among the plurality of predicted diseases.

[0161] The probabilities of various predicted diseases can be calculated using the entity prediction model mentioned above.

[0162] Step 307 , outputting non-common query information corresponding to a predicted disease in order from high to low probability.

[0163] Step 308: After receiving the user's input information for the non-common information, extract the input information to obtain supplementary information.

[0164] Step 309, determine whether the supplementary information is complete. If yes, proceed to step 310; otherwise, return to step 307.

[0165] Step 310: Determine and output the medical consultation result based on the complete supplementary information.

[0166] Step 311, querying the missing information in the symptom information.

[0167] In this embodiment, after searching for multiple predicted diseases, non-common inquiry information of these predicted diseases is obtained, and the probability of each predicted disease among the multiple predicted diseases is determined, and then the non-common inquiry information corresponding to a predicted disease is output in order from high to low according to the probability until complete supplementary information is obtained. After obtaining the complete supplementary information, the final disease is determined using the supplementary information, and then the corresponding consultation result is obtained.

[0168] It should be noted that the inquiry information in the medical knowledge graph can be information that can be used to assist in determining the disease, such as medical history, type of work, and living habits. The inquiry results can be information that is helpful in treating or alleviating the user's symptoms, such as the confirmed disease, the department visited, the items that need to be checked, medication suggestions, and living habits suggestions.

[0169] To this end, in the above embodiments, the inquiry information and the inquiry result can be directly output, or they can be processed first to make the output content more natural dialogue-like. Specifically, the inquiry information and the inquiry result can be converted into natural language text, and then the natural language text can be output.

[0170] Taking the query information as an example, the query information can be converted into natural language text in any of the following ways:

[0171] 1) Filling the query information into a preset template to obtain a natural language text;

[0172] 2) Inputting the query information into a pre-established query model, and obtaining a natural language text according to the output of the query model.

[0173] Likewise, the manner of converting the medical consultation results into natural language text is the same as above and will not be described in detail here.

[0174] In addition, it should be noted that in another embodiment of the present invention, when the consultation result is output, the user may be prompted to provide feedback on whether the output consultation result is satisfactory, such as by prompting the user through a selection control, etc. If satisfactory feedback information is received from the user, the consultation is terminated; if unsatisfactory feedback information is received from the user, the user may be guided to input more information.

[0175] The automatic medical consultation method provided by the embodiment of the present invention extracts information from the input information of the user to obtain symptom information after receiving the input information of the user. If the extracted symptom information is incomplete, inquiries are made for the missing information therein, and the information is supplemented according to the user's answer until the complete symptom information is obtained. For the complete symptom information, the pre-established medical knowledge graph is used to search for possible predicted diseases, and then the corresponding medical consultation results or inquiry information are obtained according to the predicted diseases, and the medical consultation results are determined and output according to the inquiry information.

[0176] Accordingly, an embodiment of the present invention also provides an automatic medical consultation device, such as Figure 4 Shown is a structural block diagram of the device.

[0177] In this embodiment, the device includes the following modules:

[0178] Receiving module 401, used to receive user input information;

[0179] An information extraction module 402 is used to extract information from the input information to obtain symptom information;

[0180] The judgment module 403 is used to determine whether the symptom information is complete; specifically, the judgment can be made based on whether the extracted symptom information can determine a unique disease type; if so, the symptom information is determined to be complete; otherwise, the symptom information is determined to be incomplete;

[0181] A feedback module 404 is used to inquire about the missing information in the symptom information when the judgment module 403 determines that the symptom information is incomplete, and to complete the information according to the user's answer until the complete symptom information is obtained;

[0182] Prediction module 405, used to determine the predicted disease according to the complete symptom information;

[0183] The information acquisition module 406 is used to acquire the inquiry result when there is only one predicted disease; when there are multiple predicted diseases, acquire the inquiry information corresponding to the predicted diseases;

[0184] The output module 407 is used to output the medical inquiry result or the inquiry information.

[0185] In the embodiment of the present invention, the disease information includes user information and disease condition information, such as gender, age, symptoms, medical history and other information.

[0186] As mentioned above, when the judgment module 403 cannot determine a unique disease type, it indicates that the disease information is incomplete, and there may be multiple disease types determined at this time. Accordingly, when the judgment module 403 determines that the disease information is incomplete, the feedback module 404 needs to inquire about the missing information in the disease information.

[0187] In a specific embodiment, the feedback module 404 may determine the content to be inquired based on the scores of each symptom of these diseases.

[0188] Accordingly, in this embodiment, a symptom acquisition module and a score determination module may be provided, wherein:

[0189] The symptom acquisition module is used to acquire all symptoms of the candidate disease determined according to the symptom information;

[0190] The score determination module is used to determine the discrimination and weight of each symptom, and calculate the score of the symptom based on the discrimination and weight of the symptom; the calculation of the discrimination and weight of each symptom can refer to the description in the previous embodiment of the method of the present invention, and will not be repeated here.

[0191] Accordingly, the feedback module 404 selects the symptom with the highest score from all the symptoms for inquiry. Of course, the symptom information with the highest score can be directly output, or the symptom information can be filled into a corresponding template to obtain and output a natural language text, etc., which is not limited in this embodiment of the present invention.

[0192] In some cases, the information extraction module 402 may not extract any symptom information from the user's first round of input information. In view of this situation, in another embodiment of the device of the present invention, a guidance module (not shown) may also be provided to guide the user to re-enter when the information extraction module 402 extracts no symptom information from the input information. For example: prompt the user to re-enter the details; or ask the user a question so that the user re-enters according to the question content. Of course, in actual applications, the guidance module may also be integrated with the feedback module 404, which is not limited to this embodiment of the present invention.

[0193] The prediction module 405 specifically determines the disease corresponding to the complete symptom information, that is, the predicted disease, by searching the pre-established medical knowledge graph. The medical knowledge graph has been described in detail above and will not be repeated here.

[0194] Since some different diseases may have some of the same symptoms, there may be one or more predicted diseases obtained by searching the medical knowledge graph based on the complete symptom information. To this end, in a specific implementation, the prediction module 405 includes: a graph query module, which is used to search a pre-established medical knowledge graph based on the complete symptom information to determine the predicted disease. In another specific implementation, the prediction module 405 may not only include the above-mentioned graph query module, but also include a filtering module, which is used to select a predicted disease from the multiple predicted diseases using a pre-established entity prediction model. Specifically, the entity prediction model is used to calculate the probability of each disease, and then the one with the highest probability is used as the predicted disease found.

[0195] Accordingly, the information acquisition module 406 can specifically acquire all query information of the multiple predicted diseases from the medical knowledge graph, and determine the non-common query information corresponding to each predicted disease.

[0196] It should be noted that, in one embodiment, the above-mentioned filtering module can be used to select the one with the highest probability from multiple predicted diseases as the predicted disease found. In other implementations, further inquiry can be performed, that is, the output module 407 outputs inquiry information, and the final disease is determined based on the user's input information for the inquiry information.

[0197] The output module 407 may output the query information in a variety of ways, such as:

[0198] In one embodiment, the output module 407 simultaneously outputs non-common query information corresponding to each predicted disease.

[0199] Accordingly, in this embodiment, the information extraction module 402 is further configured to extract information from the input information of the user for the non-common query information after the receiving module 401 receives the input information. The judging module 403 is further configured to determine whether the auxiliary information is complete.

[0200] Correspondingly, the prediction module 405 is also used to determine the final disease according to the complete auxiliary information when the judgment module 403 determines that the auxiliary information is complete.

[0201] Correspondingly, the information acquisition module 406 is also used to acquire the diagnosis result corresponding to the final disease. The output module 407 is also used to output the diagnosis result.

[0202] In this embodiment, when the judgment module 403 determines that the auxiliary information is incomplete, the aforementioned guidance module may guide the user to re-enter the auxiliary information.

[0203] In one embodiment, the device further includes: a probability calculation module (not shown) for determining the probability of each predicted disease among the plurality of predicted diseases. Accordingly, the output module 407 outputs the non-common query information corresponding to a predicted disease in order from high to low probability.

[0204] Accordingly, in this embodiment, the information extraction module 402 is further configured to extract information from the input information to obtain supplementary information after the receiving module 401 receives the user's input information for the non-common information. The judgment module 403 is further configured to determine whether the supplementary information is complete. The output module 407 is further configured to continue to output the next non-common inquiry information for predicting a disease when the judgment module 403 determines that the supplementary information is incomplete until all non-common inquiry information is output.

[0205] Correspondingly, the prediction module 405 is also used to determine the final disease according to the complete supplementary information after obtaining the complete supplementary information.

[0206] Correspondingly, the information acquisition module 406 is also used to acquire the diagnosis result corresponding to the final disease. The output module 407 is also used to output the diagnosis result.

[0207] It should be noted that the inquiry information in the medical knowledge graph can be information that can be used to assist in determining the disease, such as medical history, type of work, and living habits. The inquiry results can be information that is helpful in treating or alleviating the user's symptoms, such as the confirmed disease, the department visited, the items that need to be checked, medication suggestions, and living habits suggestions.

[0208] Therefore, in practical applications, the output module 407 may directly output the inquiry information and the inquiry result, or process them first to make the output content more natural and conversational.

[0209] The output module 407 may specifically include the following units:

[0210] A text conversion unit, used for converting the inquiry information or the medical inquiry result into a natural language text;

[0211] A text output unit is used to output the natural language text.

[0212] Wherein, the text conversion unit may include:

[0213] A template filling unit, used to fill the inquiry information or the medical inquiry result into a preset template to obtain a natural language text; or

[0214] The text generation unit is used to input the inquiry information or the medical inquiry result into a pre-established inquiry model, and obtain a natural language text according to the output of the inquiry model.

[0215] The automatic medical consultation device provided in the embodiment of the present invention extracts information from the input information after receiving the user's input information to obtain symptom information. If the extracted symptom information is incomplete, inquiries are made for the missing information, and the information is supplemented according to the user's answer until complete symptom information is obtained. For the complete symptom information, a pre-established medical knowledge graph is used to search for possible predicted diseases, and then the corresponding medical consultation results or inquiry information are obtained according to the predicted diseases, and the medical consultation results are determined and output according to the inquiry information.

[0216] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0217] It should be noted that the method and apparatus of the embodiments of the present invention can be applied to various terminal devices, such as mobile phones, computers, notebooks and other devices.

[0218] Figure 5 8 is a block diagram of an apparatus 800 for an automatic medical consultation method according to an exemplary embodiment. For example, the apparatus 800 may be a mobile phone, a computer, a digital broadcast terminal, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, etc.

[0219] Reference Figure 5 , the device 800 may include one or more of the following components: a processing component 802 , a memory 804 , a power component 806 , a multimedia component 808 , an audio component 810 , an input / output (I / O) interface 812 , a sensor component 814 , and a communication component 816 .

[0220] The processing component 802 generally controls the overall operation of the device 800, such as operations associated with display, phone calls, data communications, camera operations, and recording operations. The processing component 802 may include one or more processors 820 to execute instructions to complete all or part of the steps of the above-mentioned method. In addition, the processing component 802 may include one or more modules to facilitate the interaction between the processing component 802 and other components. For example, the processing component 802 may include a multimedia module to facilitate the interaction between the multimedia component 808 and the processing component 802.

[0221] The memory 804 is configured to store various types of data to support operations on the device 800. Examples of such data include instructions for any application or method operating on the device 800, contact data, phone book data, messages, pictures, videos, etc. The memory 804 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0222] The power component 806 provides power to the various components of the device 800. The power component 806 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the device 800.

[0223] The multimedia component 808 includes a screen that provides an output interface between the device 800 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundaries of the touch or slide action, but also detect the duration and pressure associated with the touch or slide operation. In some embodiments, the multimedia component 808 includes a front camera and / or a rear camera. When the device 800 is in an operating mode, such as a shooting mode or a video mode, the front camera and / or the rear camera may receive external multimedia data. Each front camera and rear camera may be a fixed optical lens system or have a focal length and optical zoom capability.

[0224] The audio component 810 is configured to output and / or input audio signals. For example, the audio component 810 includes a microphone (MIC), and when the device 800 is in an operating mode, such as a call mode, a recording mode, and a speech recognition mode, the microphone is configured to receive an external audio signal. The received audio signal can be further stored in the memory 804 or sent via the communication component 816. In some embodiments, the audio component 810 also includes a speaker for outputting audio signals.

[0225] I / O interface 812 provides an interface between processing component 802 and peripheral interface modules, such as keyboards, click wheels, buttons, etc. These buttons may include but are not limited to: home button, volume button, start button, and lock button.

[0226] The sensor assembly 814 includes one or more sensors for providing various aspects of status assessment for the device 800. For example, the sensor assembly 814 can detect the open / closed state of the device 800, the relative positioning of components, such as the display and keypad of the device 800, and the sensor assembly 814 can also detect the position change of the device 800 or a component of the device 800, the presence or absence of user contact with the device 800, the orientation or acceleration / deceleration of the device 800, and the temperature change of the device 800. The sensor assembly 814 may include a proximity sensor configured to detect the presence of a nearby object without any physical contact. The sensor assembly 814 may also include an optical sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 814 may also include an accelerometer, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

[0227] The communication component 816 is configured to facilitate wired or wireless communication between the device 800 and other devices. The device 800 can access a wireless network based on a communication standard, such as WiFi, 2G or 3G, or a combination thereof. In an exemplary embodiment, the communication component 816 receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 816 also includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.

[0228] In an exemplary embodiment, the apparatus 800 may be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to perform the above method.

[0229] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 804 including instructions, and the instructions can be executed by the processor 820 of the device 800 to complete the above-mentioned key mis-touch error correction method. For example, the non-transitory computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.

[0230] A non-temporary computer-readable storage medium, when the instructions in the storage medium are executed by the processor of a mobile terminal, enables the mobile terminal to execute a method for correcting accidental key presses, the method comprising: during the user input process, obtaining the pressing information when each key is triggered; determining the mistakenly triggered key based on the obtained pressing information; correcting the mistakenly triggered key; and determining each candidate word corresponding to the complete input string after correction.

[0231] Figure 6 1 is a schematic diagram of the structure of the server in an embodiment of the present invention. The server 1900 may have relatively large differences due to different configurations or performances, and may include one or more central processing units (CPU) 1922 (for example, one or more processors) and memory 1932, and one or more storage media 1930 (for example, one or more mass storage devices) storing application programs 1942 or data 1944. Among them, the memory 1932 and the storage medium 1930 can be short-term storage or permanent storage. The program stored in the storage medium 1930 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations on the server. Furthermore, the central processing unit 1922 can be configured to communicate with the storage medium 1930 and execute a series of instruction operations in the storage medium 1930 on the server 1900.

[0232] The server 1900 may also include one or more power supplies 1926, one or more wired or wireless network interfaces 1950, one or more input and output interfaces 1958, one or more keyboards 1956, and / or, one or more operating systems 1941, such as Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM, etc.

[0233] A non-temporary computer-readable storage medium, when the instructions in the storage medium are executed by a processor of a device, enables the device to execute the above-mentioned key mis-touch correction method.

[0234] Those skilled in the art will readily appreciate other embodiments of the present invention after considering the specification and practicing the invention disclosed herein. The present invention is intended to cover any variations, uses or adaptations of the present invention that follow the general principles of the present invention and include common knowledge or customary techniques in the art that are not disclosed in this disclosure. The description and examples are to be considered exemplary only, and the true scope and spirit of the present invention are indicated by the following claims.

[0235] It should be understood that the present invention is not limited to the precise structure described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present invention is limited only by the appended claims.

[0236] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An automatic medical consultation method, characterized in that: The method comprises: Receive user input information; Extracting information from the input information to obtain symptom information; the symptom information is relevant information in the input information required for disease diagnosis; Determining whether the symptom information is complete includes: determining whether a unique disease type can be determined according to the symptom information; if the corresponding symptom type is determined to be the unique disease type according to the symptom information, determining that the symptom information is complete; otherwise, determining that the symptom information is incomplete; If not, inquire about the missing information in the symptom information, and complete the information according to the user's answer until the symptom information whose corresponding symptom type is a unique disease type is obtained; Determine the predicted disease according to the disease information that the corresponding disease type is a unique disease type; If there is only one predicted disease, obtaining and outputting the consultation result; If there are multiple predicted diseases, obtaining query information corresponding to the predicted diseases and outputting the query information; Wherein, the method further comprises: Obtaining all symptoms of the candidate disease determined according to the symptom information; Determine the discrimination degree and weight of each symptom, and calculate the score of the symptom according to the discrimination degree and weight of the symptom; The querying of missing information in the symptom information includes: Select the highest scoring symptom for inquiry; In the case where there are multiple predicted diseases, the method further includes: A predicted disease is selected from the multiple predicted diseases using a pre-established entity prediction model; the entity prediction model is used to calculate the probability of each predicted disease, and the one with the highest probability is used as the predicted disease found; the entity prediction model is trained through existing medical question and answer data.

2. The method according to claim 1, characterized in that The method further comprises: If no symptom information is obtained by extracting the input information, the user is guided to re-enter the information.

3. The method according to claim 2, characterized in that The guiding the user to re-enter comprises: Prompt the user to re-enter details; or Ask the user a question to make the user re-enter the information based on the question.

4. The method according to claim 1, characterized in that: The determining of the predicted disease according to the symptom information that the corresponding symptom type is the only disease type includes: The pre-established medical knowledge graph is searched based on the symptom information of the unique disease type corresponding to the symptom type to determine the predicted disease.

5. The method according to claim 4, characterized in that The obtaining of inquiry information corresponding to the predicted disease includes: All query information of the multiple predicted diseases is obtained from the medical knowledge graph, and non-common query information corresponding to each predicted disease is determined.

6. The method according to claim 5, characterized in that Outputting the query information includes: simultaneously outputting non-common query information corresponding to each predicted disease; The method further comprises: After receiving the user's input information for the non-common query information, extracting the input information to obtain auxiliary information; Determining whether the auxiliary information is complete; If yes, determining and outputting the consultation result according to the complete auxiliary information; Otherwise, prompt the user to re-enter.

7. The method according to claim 5, characterized in that The outputting the inquiry information comprises: Determining the probability of each predicted disease among the plurality of predicted diseases; Output non-common query information corresponding to a predicted disease in order from high to low probability; The method further comprises: After receiving the user's input information for the non-common information, extracting the input information to obtain supplementary information; Determining whether the supplemental information is complete; If the supplementary information is incomplete, continue to output the next non-common inquiry information of the predicted disease until all the non-common inquiry information is output; After the complete supplementary information is obtained, the medical consultation result is determined and outputted according to the complete supplementary information.

8. The method according to claim 1, characterized in that The output of the medical consultation results includes: Converting the medical consultation results into natural language text; The natural language text is output.

9. The method according to claim 8, characterized in that The converting of the medical consultation result into natural language text comprises: Filling the medical consultation result into a preset template to obtain a natural language text; or The inquiry result is input into a pre-established inquiry model, and a natural language text is obtained according to the output of the inquiry model.

10. An automatic medical consultation device, characterized in that: The device comprises: A receiving module, used for receiving user input information; An information extraction module is used to extract information from the input information to obtain symptom information; the symptom information is relevant information in the input information that is required for disease judgment; A judgment module is specifically used to determine whether a unique disease type can be determined according to the symptom information; if the corresponding symptom type is determined to be the unique disease type according to the symptom information, then the symptom information is determined to be complete; otherwise, the symptom information is determined to be incomplete; A feedback module, used for, when the judgment module determines that the symptom information is incomplete, inquiring about the missing information in the symptom information, and completing the information according to the user's answer, until the symptom information whose corresponding symptom type is a unique disease type is obtained; A prediction module, used for determining a predicted disease according to the symptom information whose corresponding symptom type is a unique disease type; An information acquisition module, used to obtain the consultation result when there is only one predicted disease; and to obtain the inquiry information corresponding to the predicted disease when there are multiple predicted diseases; An output module, used for outputting the medical inquiry result or the inquiry information; Wherein, the device further comprises: A symptom acquisition module, used to acquire all symptoms of the candidate disease determined according to the symptom information; A score determination module, used to determine the discrimination and weight of each symptom, and calculate the score of the symptom according to the discrimination and weight of the symptom; The feedback module selects the symptom with the highest score to inquire about; In the case where there are multiple predicted diseases, the device further comprises: A filtering module is used to select a predicted disease from the multiple predicted diseases using a pre-established entity prediction model; the entity prediction model is used to calculate the probability of each predicted disease, and the one with the highest probability is used as the predicted disease found; the entity prediction model is trained through existing medical question and answer data.

11. The device according to claim 10, characterized in that The device also includes: The guiding module is used to guide the user to re-input the information when the information extraction module fails to obtain the symptom information after extracting the input information.

12. The device according to claim 11, characterized in that The guiding module is specifically used to guide the user to re-enter in the following ways: prompting the user to re-enter detailed content; or asking questions to the user so that the user re-enters according to the question content.

13. The device according to claim 11, characterized in that The prediction module includes: a graph query module, which is used to search a pre-established medical knowledge graph according to the symptom information of a unique disease type corresponding to the symptom type, and determine the predicted disease.

14. The device according to claim 13, characterized in that The information acquisition module is specifically used to obtain all query information of the multiple predicted diseases from the medical knowledge graph, and determine the non-common query information corresponding to each predicted disease.

15. The device according to claim 14, characterized in that The output module is further used to simultaneously output non-common query information corresponding to each predicted disease; The information extraction module is further configured to extract information from the input information of the user for the non-common inquiry information after the receiving module receives the input information of the user to obtain auxiliary information; The judging module is further used to determine whether the auxiliary information is complete; The prediction module is further configured to determine a final disease according to the complete auxiliary information when the judgment module determines that the auxiliary information is complete; The information acquisition module is further used to obtain the diagnosis result corresponding to the final disease; The output module is further used to output the consultation result; The guiding module is further used to guide the user to re-enter the auxiliary information when the judging module determines that the auxiliary information is incomplete.

16. The device according to claim 14, characterized in that The device also includes: A probability calculation module, used to determine the probability of each predicted disease among the multiple predicted diseases; The output module is also used to output non-common query information corresponding to a predicted disease in order from high to low probability; The information extraction module is further configured to extract information from the input information of the user for the non-common information after the receiving module receives the input information of the user for the non-common information to obtain supplementary information; The judging module is further used to determine whether the supplementary information is complete; The output module is further configured to continue outputting the next non-common inquiry information of the predicted disease when the judgment module determines that the supplementary information is incomplete, until all the non-common inquiry information is outputted; The prediction module is further used to determine the final disease according to the complete supplementary information after obtaining the complete supplementary information; The information acquisition module is further used to obtain the diagnosis result corresponding to the final disease; The output module is also used to output the medical consultation result.

17. The device according to claim 10, characterized in that The output module comprises: A text conversion unit, used for converting the inquiry information or the medical inquiry result into a natural language text; A text output unit is used to output the natural language text.

18. The device according to claim 17, characterized in that The text conversion unit comprises: A template filling unit, used to fill the inquiry information or the medical inquiry result into a preset template to obtain a natural language text; or The text generation unit is used to input the inquiry information or the medical inquiry result into a pre-established inquiry model, and obtain a natural language text according to the output of the inquiry model.

19. A computer device, characterized in that: include: One or more processors, memory; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the method according to any one of claims 1 to 9.

20. A readable storage medium having instructions stored thereon, wherein the instructions are executed to implement the method according to any one of claims 1 to 9.

Images