CN116313162B - Medical inquiry system based on AI model - Google Patents
Medical inquiry system based on AI model Download PDFInfo
- Publication number
- CN116313162B CN116313162B CN202310530384.7A CN202310530384A CN116313162B CN 116313162 B CN116313162 B CN 116313162B CN 202310530384 A CN202310530384 A CN 202310530384A CN 116313162 B CN116313162 B CN 116313162B
- Authority
- CN
- China
- Prior art keywords
- medical
- text
- evaluation parameter
- similarity
- similarity evaluation
- 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
- 238000011156 evaluation Methods 0.000 claims abstract description 186
- 230000004044 response Effects 0.000 claims abstract description 173
- 208000024891 symptom Diseases 0.000 claims abstract description 115
- 238000001514 detection method Methods 0.000 claims abstract description 76
- 238000000034 method Methods 0.000 claims abstract description 16
- 230000008569 process Effects 0.000 claims abstract description 14
- 238000012216 screening Methods 0.000 claims description 20
- 238000003745 diagnosis Methods 0.000 claims description 18
- 239000013598 vector Substances 0.000 claims description 12
- 230000011218 segmentation Effects 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 8
- 238000002372 labelling Methods 0.000 claims description 7
- 201000010099 disease Diseases 0.000 claims description 4
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 claims description 4
- 230000006872 improvement Effects 0.000 abstract description 8
- 238000010586 diagram Methods 0.000 description 9
- 239000003814 drug Substances 0.000 description 5
- 229940079593 drug Drugs 0.000 description 5
- 238000007689 inspection Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 206010000496 acne Diseases 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000008280 blood Substances 0.000 description 1
- 210000004369 blood Anatomy 0.000 description 1
- 230000036772 blood pressure Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 150000002632 lipids Chemical class 0.000 description 1
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
- G16H80/00—ICT specially adapted for facilitating communication between medical practitioners or patients, e.g. for collaborative diagnosis, therapy or health monitoring
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation or dialogue systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/334—Query execution
- G06F16/3344—Query execution using natural language analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/205—Parsing
- G06F40/216—Parsing using statistical methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/253—Grammatical analysis; Style critique
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Databases & Information Systems (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Medical Informatics (AREA)
- Human Computer Interaction (AREA)
- Public Health (AREA)
- Primary Health Care (AREA)
- Epidemiology (AREA)
- Pathology (AREA)
- Probability & Statistics with Applications (AREA)
- Biomedical Technology (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention relates to the technical field of medical inquiry systems, in particular to a medical inquiry system based on an AI model, which comprises: the inquiry module is used for carrying out text dialogue inquiry; the cloud database is connected with the inquiry module and used for storing data updated in real time by the cloud and corresponding inquiry data of a plurality of medical inquiry processes; the detection module is respectively connected with the inquiry module and the cloud database; the central control module is used for judging whether to call a fixed medical response text in the cloud database according to the patient symptom similarity evaluation parameters detected by the first similarity detection unit, judging whether the accuracy of the medical response text is in an allowable range according to the medical response text similarity evaluation parameters detected by the second similarity detection unit, and adjusting the standard medical feature recognition granularity to a corresponding granularity; the invention realizes the improvement of the accuracy of inquiry.
Description
Technical Field
The invention relates to the technical field of medical consultation, in particular to a medical consultation system based on an AI model.
Background
The medical inquiry system in the prior art has the problem that the inquiry precision is influenced due to the fact that symptom text input is not standard and the content similarity detection of medical response text is inaccurate.
Chinese patent publication No.: CN114220537a discloses an AI intelligent on-line diagnosis method and cloud system based on internet hospitals, comprising: constructing a human symptom information base according to a human body structure diagram; acquiring registration information corresponding to a current registrant of a current target online consultation platform; collecting current symptom information and uploaded inspection report information input by the target analyst; basic information and historical diagnosis information corresponding to each registered doctor of the on-target-line diagnosis and inquiry platform are obtained; according to the current symptom information input by the target analyst and the uploaded inspection report information, confirming an adaptation registration doctor corresponding to the target analyst; therefore, the intelligent online AI diagnosis method and cloud system based on the Internet hospital have the following problems: the accuracy of the interrogation system is affected by the lack of standardization of symptom characteristics in the interrogation process and the passivity of the medical response process.
Disclosure of Invention
In order to solve the problem that in the prior art, the symptom characteristics in the inquiry process are not standard and the passivity of the medical response process affects the accuracy of the inquiry system, the invention provides a medical inquiry system based on an AI model, which comprises the following steps: the inquiry module is used for carrying out text dialogue inquiry and comprises a medical inquiry unit used for inquiring medical questions and a response unit connected with the medical inquiry unit and used for responding the medical questions; the cloud database is connected with the inquiry module and used for storing data updated in real time by the cloud and corresponding inquiry data of a plurality of medical inquiry processes; the detection module is respectively connected with the inquiry module and the cloud database and comprises a first similarity detection unit which is respectively connected with the medical inquiry unit and the cloud database and used for detecting and calculating the similarity evaluation parameters of the symptoms of the patient, a second similarity detection unit which is respectively connected with the response unit and the cloud database and used for detecting and calculating the similarity evaluation parameters of the medical response text, and a screening unit which is respectively connected with the medical inquiry unit and the response unit and used for screening out the corresponding symptom text of the patient and the corresponding medical response text, wherein the similarity evaluation parameters of the symptoms of the patient are similarity evaluation parameters of the symptom text of the patient meeting the condition of the first symptom feature quantity and standard symptom sentences stored in the cloud database, and the similarity evaluation parameters of the medical response text are similarity evaluation parameters of the medical response text output by the response unit and the standard diagnosis sentences stored in the cloud database; the central control module is respectively connected with the inquiry module, the cloud database and the detection module and is used for judging whether to call a fixed medical response text in the cloud database according to the patient symptom similarity evaluation parameter detected by the first similarity detection unit and judging whether the accuracy of the medical response text is in an allowable range according to the medical response text similarity evaluation parameter detected by the second similarity detection unit and adjusting the standard medical feature recognition granularity to a corresponding granularity; the first symptom characteristic quantity condition is that the symptom characteristic quantity in the patient symptom text output by the medical questioning unit is larger than the preset symptom characteristic quantity.
Further, the central control module judges whether to call the fixed medical response text in the cloud database according to the patient symptom similarity evaluation parameters detected by the first similarity detection unit,
if the patient symptom similarity evaluation parameter is smaller than the preset symptom similarity evaluation parameter, the central control module judges that the complexity of the patient symptom text output by the medical questioning unit is beyond the allowable range, and the central control module retrieves the fixed medical response text in the cloud database and sends the fixed medical response text to the response unit after the retrieval is completed;
and if the patient symptom similarity evaluation parameter is greater than or equal to the preset symptom similarity evaluation parameter, the central control module judges that the complexity of the patient symptom text is within the allowable range and controls the response unit to output the corresponding medical response text.
Further, the central control module determines whether the accuracy of the medical response is within the allowable range according to the calculated similarity evaluation parameter of the medical response text detected by the second similarity detection unit, wherein,
the first type of judgment mode is that the central control module judges that the accuracy of the medical response is lower than an allowable range under the condition of presetting a first similarity evaluation parameter, and adjusts the standard medical feature identification granularity to a corresponding granularity by calculating the difference value between the similarity evaluation parameter of the medical response text and the similarity evaluation parameter of the preset response text;
The second type of judgment mode is that the central control module judges that the accuracy of the medical response is within an allowable range under the condition of presetting a second similarity evaluation parameter;
the condition of the preset first similarity evaluation parameter is that the medical response text similarity evaluation parameter is smaller than or equal to the preset response text similarity evaluation parameter; the condition of the preset second similarity evaluation parameter is that the medical response text similarity evaluation parameter is larger than the preset response text similarity evaluation parameter.
Further, the central control module determines three types of adjustment modes aiming at standard medical feature recognition granularity according to the difference value of the medical response text similarity evaluation parameter and the preset response text similarity evaluation parameter under the condition of the preset first similarity evaluation parameter, wherein,
the first type of adjustment mode is that the central control module adjusts the standard medical feature identification granularity to a preset identification granularity under the condition of presetting a first similarity evaluation parameter difference value;
the second type of adjustment mode is that the central control module adjusts the standard medical feature identification granularity to the first identification granularity by using a preset second granularity adjustment coefficient under the condition of presetting a second similarity evaluation parameter difference value;
The third type of adjustment mode is that the central control module adjusts the standard medical characteristic identification granularity to the second identification granularity by using a preset first granularity adjustment coefficient under the condition of presetting a third similarity evaluation parameter difference value;
the difference value condition of the preset first similarity evaluation parameter is that the difference value of the similarity evaluation parameter of the medical response text and the similarity evaluation parameter of the preset response text is smaller than or equal to the difference value of the similarity evaluation parameter of the preset first response sentence; the difference value condition of the preset second similarity evaluation parameter is that the difference value of the medical response text similarity evaluation parameter and the preset response text similarity evaluation parameter is larger than the difference value of the preset first response sentence similarity evaluation parameter and smaller than or equal to the difference value of the preset second response sentence similarity evaluation parameter; the difference value condition of the preset third similarity evaluation parameter is that the difference value between the medical response text similarity evaluation parameter and the preset response text similarity evaluation parameter is larger than the difference value of the preset second response sentence similarity evaluation parameter; the preset first answer sentence similarity evaluation parameter difference value is smaller than the preset second answer sentence similarity evaluation parameter difference value, and the preset first granularity adjustment coefficient is smaller than the preset second granularity adjustment coefficient.
Further, the standard medical feature identification granularity is granularity of the feature aiming at the disease symptom in the re-questioning notification sent by the system to the questioning patient.
Further, the screening unit includes:
an interception component connected with the medical question unit and the response unit respectively, for intercepting the patient symptom text characteristic text and the medical response characteristic text by using an H00K hook;
the label generation component is connected with the interception component and is used for carrying out labeling processing and storage on the intercepted medical feature texts through jieba segmentation, wherein the medical feature texts comprise patient symptom feature texts and medical response feature texts;
wherein the tag generation component comprises:
the paragraph structure serialization plug-in is used for carrying out serialization operation on the medical feature text according to the number of characters between adjacent commas of the medical feature text and respectively transmitting the medical feature text after serialization to the first similarity detection unit and the second similarity detection unit;
the vocabulary diversity serialization plug-in is used for carrying out serialization operation on the medical feature texts according to the number of different vocabularies and the types of the vocabularies and respectively transmitting the serialized medical feature texts to the first similarity detection unit and the second similarity detection unit;
The grammar complexity plug-in is used for carrying out serialization operation on the medical feature text according to the grammar complexity evaluation parameters of the words and respectively transmitting the serialized medical feature text to the first similarity detection unit and the second similarity detection unit;
the vocabulary structure plug-in is used for carrying out serialization operation on the medical feature text according to whether the vocabulary is an overlapping word or not and respectively transmitting the serialized medical feature text to the first similarity detection unit and the second similarity detection unit;
a keyword fingerprint plug-in for storing the medical feature text after the removed stop word;
for the paragraph structure serialization plugin, if the number of characters between adjacent commas is an integer multiple of a preset character number unit threshold value which is larger than 0, the paragraph structure serialization plugin takes the number of characters between the current adjacent commas as a vector standard for serialization;
the tag generation component ranks and stores punctuation marks by rank as the medical feature text is serialized using different serialization plugins.
Further, the interception component judges whether to intercept the medical feature text according to the number of repeated words in the medical feature text,
If the number of repeated words in the medical feature text is greater than or equal to the standard number of words in the medical feature text, the interception component judges that the medical feature text is intercepted;
the calculation formula of the repeated word number ratio in the medical feature text is as follows:
wherein G is the number of repeated words in the medical feature text to be counted, M is the total number of repeated words in the medical feature text, M Total (S) Is the total number of words of the medical feature text.
Further, the screening unit judges whether the intercepted text is a valid intercepted text according to the character set length of the intercepted medical feature text when the intercepting component completes the interception of the corresponding medical feature text,
if the character set length of the intercepted medical feature text is smaller than the preset character set length, the screening unit judges that the intercepted text is not an effective intercepted text, and uses the HOOK to remove the intercepted text;
if the character set length of the intercepted medical feature text is larger than the preset character set length, the screening unit judges that the intercepted text is a valid text, and performs part-of-speech tagging on the intercepted text by using jieba segmentation.
Further, the calculation formula of the complexity evaluation parameter of the grammar of the words is as follows:
Wherein S is the complexity evaluation parameter of the grammar of the word, A is the number of misplaced Chinese characters of the grammar components, a is the weight coefficient of the number of misplaced Chinese characters of the grammar components, B is the number of grammar components, B is the weight coefficient of the number of grammar components, C is the average Chinese character number among grammar components, and C is the weight coefficient of the average Chinese character number among grammar components; a=0.4, b=0.2, c=0.4.
Further, the calculation formula of the patient symptom similarity evaluation parameter or the medical response text similarity evaluation parameter is as follows:
wherein U is a patient symptom similarity evaluation parameter or a medical response text similarity evaluation parameter, E is a cosine similarity between a patient symptom text satisfying a first symptom feature quantity condition and standard symptom sentences stored in the cloud database or a cosine similarity between a medical response text and standard diagnostic sentences stored in the cloud database, E is a cosine similarity weight coefficient, P is an edit distance between a patient symptom text satisfying a first symptom feature quantity condition and standard symptom sentences stored in the cloud database or an edit distance between a medical response text and standard diagnostic sentences stored in the cloud database, P is an edit distance weight coefficient, e+p=1;
When the medical feature text similarity is judged by taking the absolute difference as a standard, the central control module increases the editing distance weight coefficient so that p is more than e;
and when the medical feature text similarity determination is performed by taking the relative difference as a standard, the central control module adjusts the cosine similarity weight coefficient to be higher so that e is more than p.
Compared with the prior art, the system has the beneficial effects that through the arranged inquiry module, cloud database, detection module and central control module, the corresponding characteristic texts are intercepted and the intercepted texts are subjected to labeling processing and storage through the arranged interception component and the label generation component, so that the influence on the accuracy of similarity judgment of the medical response texts output by the response unit and standard diagnosis sentences due to inaccuracy of text labeling processing is reduced; the medical text similarity evaluation parameters are calculated through the set second similarity detection unit to judge the similarity degree, so that the accuracy of the output content of the response unit is improved; according to the similarity between the medical response text detected by the second similarity detection unit and the standard diagnosis sentences stored in the cloud database, whether the accuracy of the medical response text is within an allowable range is judged, the standard medical feature recognition granularity is adjusted to the corresponding granularity, the reduction of the accuracy of the inquiry caused by insufficient granularity of the inquiry text is reduced through the adjustment of the granularity, and the improvement of the accuracy of the medical inquiry is realized.
Further, according to the system, through the preset symptom similarity evaluation parameters, whether the fixed medical response text in the cloud database is called or not is judged according to the symptom similarity between the patient symptom text meeting the first symptom characteristic quantity condition and the standard symptom statement stored in the cloud database, which is detected by the first similarity detection unit, the influence on the medical inquiry accuracy due to the fact that whether the fixed medical response text in the cloud database is called or not is inaccurate is reduced, and the medical inquiry accuracy is further improved.
Further, the system of the invention determines three types of adjustment modes aiming at standard medical feature recognition granularity according to the set preset first answer sentence similarity, preset second answer sentence similarity evaluation parameter difference, preset first granularity adjustment coefficient and preset second granularity adjustment coefficient and the difference between the medical answer text similarity evaluation parameter and the preset answer text similarity evaluation parameter, thereby reducing the influence on the accuracy of the inquiry due to inaccurate adjustment of the standard medical feature recognition granularity and further realizing the improvement of the accuracy of the medical inquiry.
Further, the system of the invention carries out serialization operation on the texts according to the number of different vocabularies and the types of the vocabularies through the set vocabulary diversity serialization plugin, grammar complexity plugin and vocabulary structure plugin, and transmits the serialized texts to the second similarity detection unit, carries out serialization operation on the texts according to the complexity evaluation parameters of the grammar of the words and transmits the serialized texts to the second similarity detection unit, thereby reducing the influence of insufficient precision of text serialization on similarity detection; the serialization operation is carried out on the text according to the number of characters between adjacent commas of the sentence paragraphs, so that the serialization accuracy is improved; the keyword fingerprint plug-in arranged in the label generating assembly is used for storing the characters after the removed stop words, so that the serialization diversity is improved; through arranging and storing punctuation marks according to the bit times when different serialization standards are used for serializing the texts, the serialization accuracy is improved, and the medical inquiry accuracy is further improved.
Furthermore, the system judges whether the text is intercepted or not according to the number of the repeated words in the text by the arranged interception component, so that the influence of inaccurate setting of interception standards on text interception accuracy is reduced, and the improvement of medical inquiry accuracy is further realized.
Furthermore, the system adjusts the weight coefficient according to different standard conditions through the set adjustment process aiming at the weight coefficient, improves the accuracy of similarity detection, and further realizes the improvement of medical inquiry accuracy.
Drawings
FIG. 1 is a block diagram showing the overall structure of a medical inquiry system based on an AI model according to an embodiment of the present invention;
FIG. 2 is a block diagram of a detection module of a medical consultation system based on an AI model according to an embodiment of the present invention;
FIG. 3 is a block diagram showing a screening unit of a medical consultation system based on an AI model according to an embodiment of the present invention;
fig. 4 is a block diagram of a second similarity detection unit of the AI model-based medical inquiry system according to an embodiment of the present invention.
Detailed Description
In order that the objects and advantages of the invention will become more apparent, the invention will be further described with reference to the following examples; it should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
It should be noted that, the data in this embodiment are obtained by comprehensive analysis and evaluation according to the historical inquiry data and the corresponding detection of the previous inquiry process before the inquiry is performed by the system of the present invention; the system integrates 2432 monitored medical response text similarity evaluation parameters, patient symptom similarity evaluation parameters and the number of symptom characteristics in the patient symptom text before the inquiry and comprehensively determines the numerical value of each preset parameter standard aiming at the inquiry. It can be understood by those skilled in the art that the determining manner of the system according to the present invention for the single item of parameter may be to select the value with the highest duty ratio as the preset standard parameter according to the data distribution, so long as the system according to the present invention can clearly define different specific situations in the single item determination process through the obtained value.
Preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are merely for explaining the technical principles of the present invention, and are not intended to limit the scope of the present invention.
Furthermore, it should be noted that, in the description of the present invention, unless explicitly specified and limited otherwise, the terms "mounted," "connected," and "connected" are to be construed broadly, and may be either fixedly connected, detachably connected, or integrally connected, for example; can be mechanically or electrically connected; can be directly connected or indirectly connected through an intermediate medium, and can be communication between two elements. The specific meaning of the above terms in the present invention can be understood by those skilled in the art according to the specific circumstances.
Referring to fig. 1, fig. 2, fig. 3, and fig. 4, which are an overall block diagram, a detection module block diagram, a screening unit block diagram, and a second similarity detection unit block diagram of a medical inquiry system based on an AI model according to an embodiment of the present invention, respectively; the invention relates to a medical inquiry system based on an AI model, which comprises:
the inquiry module is used for carrying out text dialogue inquiry and comprises a medical inquiry unit used for inquiring medical questions and a response unit connected with the medical inquiry unit and used for responding the medical questions;
The cloud database is connected with the inquiry module and used for storing data updated in real time by the cloud and corresponding inquiry data of a plurality of medical inquiry processes;
the detection module is respectively connected with the inquiry module and the cloud database and comprises a first similarity detection unit which is respectively connected with the medical inquiry unit and the cloud database and used for detecting and calculating the symptom similarity evaluation parameters of the patient, a second similarity detection unit which is respectively connected with the response unit and the cloud database and used for detecting and calculating the similarity evaluation parameters of the medical response text, and a screening unit which is respectively connected with the medical inquiry unit and the response unit and used for screening out the symptom texts of the corresponding patient and the corresponding medical response texts,
the patient symptom similarity evaluation parameters are similarity evaluation parameters of a patient symptom text meeting the first symptom feature quantity condition and standard symptom sentences stored in the cloud database, and the medical response text similarity evaluation parameters are similarity evaluation parameters of a medical response text output by the response unit and standard diagnosis sentences stored in the cloud database;
the central control module is respectively connected with the inquiry module, the cloud database and the detection module and is used for judging whether to call a fixed medical response text in the cloud database according to the patient symptom similarity evaluation parameter detected by the first similarity detection unit and judging whether the accuracy of the medical response text is in an allowable range according to the medical response text similarity evaluation parameter detected by the second similarity detection unit and adjusting the standard medical feature recognition granularity to a corresponding granularity;
The first symptom characteristic quantity condition is that the symptom characteristic quantity in the patient symptom text output by the medical questioning unit is larger than the preset symptom characteristic quantity.
Specifically, the number of symptom features in the patient symptom text output by the medical question unit is denoted as Q, and the preset number of symptom features is denoted as Q0.
Specifically, preferred examples of symptom characteristics are: the throat is pimple, tone color is changed, blood pressure is changed, and blood lipid is changed, but the embodiment of symptom characteristics is not limited to the above embodiment, and other embodiments are not described herein.
According to the system, through the set inquiry module, cloud database, detection module and central control module, the corresponding characteristic text is intercepted and the intercepted text is subjected to labeling processing and storage through the set interception component and the label generation component, so that the influence on the accuracy of judging the similarity of the medical response text output by the response unit and the standard diagnosis statement due to inaccuracy of text labeling processing is reduced; the medical text similarity evaluation parameters are calculated through the set second similarity detection unit to judge the similarity degree, so that the accuracy of the output content of the response unit is improved; according to the similarity between the medical response text detected by the second similarity detection unit and the standard diagnosis sentences stored in the cloud database, whether the accuracy of the medical response text is within an allowable range is judged, the standard medical feature recognition granularity is adjusted to the corresponding granularity, the reduction of the accuracy of the inquiry caused by insufficient granularity of the inquiry text is reduced through the adjustment of the granularity, and the improvement of the accuracy of the medical inquiry is realized.
Specifically, the second similarity detection unit comprises a word segmentation component for performing word segmentation processing on two Chinese sentences, a calculation component connected with the word segmentation component for calculating cosine similarity between the two Chinese sentences and edit distances of two texts by using Levenshtein, and a weighting component connected with the calculation component for calculating patient symptom similarity evaluation parameters and medical response text similarity evaluation parameters.
Specifically, the first similarity detection unit and the second similarity detection unit have the same internal structure.
Specifically, the computing component includes:
the cosine similarity calculation plug-in is used for calculating the cosine similarity between two Chinese sentences;
and the editing distance calculating plug-in is used for calculating the editing distance of the two texts by using the Levenshtein.
Specifically, the principle of the Levenshtein to calculate the distance is: calculating the minimum number of editing operations required to edit one character string into another, for example, only 3 single character editing operations are required to modify the character string from the character string 'kitten' to the character string 'position', so that the Levenshtein distance of 'kitten' and 'position' is 3; wherein the character operation includes: deleting a character, inserting a character, and modifying a character.
It can be understood by those skilled in the art that the cosine similarity calculation and the Levenshtein calculation edit distance are common calculation means, which belong to the prior art, the cosine value of the cosine similarity calculation is essentially that the cosine value of the two vector included angles in the vector space is used as a standard for measuring the difference between two individuals, and the cosine value of the vector included angles is the prior art and will not be described herein.
With continued reference to fig. 1, the central control module determines whether to call a fixed medical response text in a cloud database according to the patient symptom similarity evaluation parameter detected by the first similarity detection unit,
if the patient symptom similarity evaluation parameter is smaller than the preset symptom similarity evaluation parameter, the central control module judges that the complexity of the patient symptom text output by the medical questioning unit is beyond the allowable range, and the central control module retrieves the fixed medical response text in the cloud database and sends the fixed medical response text to the response unit after the retrieval is completed;
and if the patient symptom similarity evaluation parameter is greater than or equal to the preset symptom similarity evaluation parameter, the central control module judges that the complexity of the patient symptom text is within the allowable range and controls the response unit to output the corresponding medical response text.
According to the system, through the preset symptom similarity evaluation parameters, whether the fixed medical response text in the cloud database is called or not is judged according to the symptom similarity between the patient symptom text meeting the first symptom characteristic quantity condition and the standard symptom statement stored in the cloud database, which is detected by the first similarity detection unit, the influence on the medical inquiry accuracy due to the fact that whether the fixed medical response text in the cloud database is called or not is judged inaccurately is reduced, and the medical inquiry accuracy is further improved.
With continued reference to fig. 1, the central control module determines whether the accuracy of the medical response is within the allowable range according to the similarity evaluation parameter of the medical response text detected and calculated by the second similarity detection unit, where,
the first type of judgment mode is that the central control module judges that the accuracy of the medical response is lower than an allowable range under the condition of presetting a first similarity evaluation parameter, and adjusts the standard medical feature identification granularity to a corresponding granularity by calculating the difference value between the similarity evaluation parameter of the medical response text and the similarity evaluation parameter of the preset response text;
The second type of judgment mode is that the central control module judges that the accuracy of the medical response is within an allowable range under the condition of presetting a second similarity evaluation parameter;
the condition of the preset first similarity evaluation parameter is that the medical response text similarity evaluation parameter is smaller than or equal to the preset response text similarity evaluation parameter; the condition of the preset second similarity evaluation parameter is that the medical response text similarity evaluation parameter is larger than the preset response text similarity evaluation parameter.
Specifically, the medical response text similarity evaluation parameter is denoted as E, the preset response text similarity evaluation parameter is denoted as E0, the difference between the medical response text similarity evaluation parameter and the preset response text similarity evaluation parameter is denoted as Δe, and Δe=e-E0 is set.
With continued reference to fig. 1, the central control module determines three types of adjustment modes for the standard medical feature recognition granularity according to the difference between the medical response text similarity evaluation parameter and the preset response text similarity evaluation parameter under the condition of the preset first similarity evaluation parameter, wherein,
the first type of adjustment mode is that the central control module adjusts the standard medical feature identification granularity to a preset identification granularity under the condition of presetting a first similarity evaluation parameter difference value;
The second type of adjustment mode is that the central control module adjusts the standard medical feature identification granularity to the first identification granularity by using a preset second granularity adjustment coefficient under the condition of presetting a second similarity evaluation parameter difference value;
the third type of adjustment mode is that the central control module adjusts the standard medical characteristic identification granularity to the second identification granularity by using a preset first granularity adjustment coefficient under the condition of presetting a third similarity evaluation parameter difference value;
the difference value condition of the preset first similarity evaluation parameter is that the difference value of the similarity evaluation parameter of the medical response text and the similarity evaluation parameter of the preset response text is smaller than or equal to the difference value of the similarity evaluation parameter of the preset first response sentence; the difference value condition of the preset second similarity evaluation parameter is that the difference value of the medical response text similarity evaluation parameter and the preset response text similarity evaluation parameter is larger than the difference value of the preset first response sentence similarity evaluation parameter and smaller than or equal to the difference value of the preset second response sentence similarity evaluation parameter; the difference value condition of the preset third similarity evaluation parameter is that the difference value between the medical response text similarity evaluation parameter and the preset response text similarity evaluation parameter is larger than the difference value of the preset second response sentence similarity evaluation parameter; the preset first answer sentence similarity evaluation parameter difference value is smaller than the preset second answer sentence similarity evaluation parameter difference value, and the preset first granularity adjustment coefficient is smaller than the preset second granularity adjustment coefficient.
Specifically, the difference of similarity evaluation parameters of the preset first answer sentence is denoted as Δe1, the difference of similarity evaluation parameters of the preset second answer sentence is denoted as Δe2, the preset first granularity adjustment coefficient is denoted as α1, the preset second granularity adjustment coefficient is denoted as α2, the preset recognition granularity is denoted as N0, Δe1 < [ Δe2 ], 1 < α1 < α2, the adjusted standard medical feature recognition granularity is denoted as N ', N' =n0× (1+αi)/2 is set, wherein αi is the preset ith granularity adjustment coefficient, and i=1, 2 is set.
According to the system, three types of adjustment modes aiming at standard medical feature recognition granularity are determined according to the difference value of the medical response text similarity evaluation parameter and the preset response text similarity evaluation parameter, so that the influence on the accuracy of the inquiry due to inaccurate adjustment of the standard medical feature recognition granularity is reduced, and the improvement of the accuracy of the medical inquiry is further realized.
With continued reference to fig. 1, the standard medical feature recognition granularity is the granularity of features for symptoms of a disease in a re-questioning notification issued by the system to a interview patient.
In particular, as a preferred embodiment of the present invention, the standard medical profile identifies a granularity as the number of symptom profiles in a symptom profile of a certain class of disease in a patient.
With continued reference to fig. 1 and 3, the screening unit includes:
an interception component connected with the medical question unit and the response unit respectively, for intercepting the patient symptom text characteristic text and the medical response characteristic text by using an H00K hook;
the label generation component is connected with the interception component and is used for carrying out labeling processing and storage on the intercepted medical feature texts through jieba segmentation, wherein the medical feature texts comprise patient symptom feature texts and medical response feature texts;
wherein the tag generation component comprises:
the paragraph structure serialization plug-in is used for carrying out serialization operation on the medical feature text according to the number of characters between adjacent commas of the medical feature text and respectively transmitting the medical feature text after serialization to the first similarity detection unit and the second similarity detection unit;
the vocabulary diversity serialization plug-in is used for carrying out serialization operation on the medical feature texts according to the number of different vocabularies and the types of the vocabularies and respectively transmitting the serialized medical feature texts to the first similarity detection unit and the second similarity detection unit;
The grammar complexity plug-in is used for carrying out serialization operation on the medical feature text according to the grammar complexity evaluation parameters of the words and respectively transmitting the serialized medical feature text to the first similarity detection unit and the second similarity detection unit;
the vocabulary structure plug-in is used for carrying out serialization operation on the medical feature text according to whether the vocabulary is an overlapping word or not and respectively transmitting the serialized medical feature text to the first similarity detection unit and the second similarity detection unit;
a keyword fingerprint plug-in for storing the medical feature text after the removed stop word;
for the paragraph structure serialization plugin, if the number of characters between adjacent commas is an integer multiple of a preset character number unit threshold value which is larger than 0, the paragraph structure serialization plugin takes the number of characters between the current adjacent commas as a vector standard for serialization;
the tag generation component ranks and stores punctuation marks by rank as the medical feature text is serialized using different serialization plugins.
Specifically, a preferred embodiment of the specific process of tag generation is to segment each sentence according to the sentence sequence in the paragraph, and sequence each sentence into a character string according to the word sequence, and finally splice the serialized character strings of all sentences together to form the serialized character string of the paragraph; converting each sentence in the paragraph into a vector, as a preferred embodiment, using Word2Vec Word vector model to effect vector conversion, and concatenating the vectors of all sentences to form a vector representation of the paragraph; for a particular type of paragraph, a particular serialization approach may be used, for example, for program code paragraphs, the code may be broken down into each line, then each line of code is serialized into a string, and the serialized strings of all lines of code are concatenated to form a serialized string of the code paragraph.
According to the system, through the vocabulary diversity serialization plugin, the grammar complexity plugin and the vocabulary structure plugin, the text is serialized according to the number of different vocabularies and the types of the vocabularies, the serialized text is conveyed to the second similarity detection unit, the text is serialized according to the complexity evaluation parameters of the grammar of the words, and the serialized text is conveyed to the second similarity detection unit, so that the influence of insufficient accuracy of text serialization on similarity detection is reduced; the serialization operation is carried out on the text according to the number of characters between adjacent commas of the sentence paragraphs, so that the serialization accuracy is improved; the keyword fingerprint plug-in arranged in the label generating assembly is used for storing the characters after the removed stop words, so that the serialization diversity is improved; through arranging and storing punctuation marks according to the bit times when different serialization standards are used for serializing the texts, the serialization accuracy is improved, and the medical inquiry accuracy is further improved.
With continued reference to fig. 1 and 2, the interception component determines whether to intercept the medical feature text based on the number of repeated words in the medical feature text,
If the number of repeated words in the medical feature text is greater than or equal to the standard number of words in the medical feature text, the interception component judges that the medical feature text is intercepted;
the calculation formula of the repeated word number ratio in the medical feature text is as follows:
wherein G is the number of repeated words in the medical feature text to be counted, M is the total number of repeated words in the medical feature text, M Total (S) Is the total number of words of the medical feature text.
According to the system, whether the text is intercepted or not is judged according to the number of the repeated words in the text by the arranged interception component, so that the influence of inaccurate setting of interception standards on text interception accuracy is reduced, and the improvement of medical inquiry accuracy is further realized.
With continued reference to fig. 1 and 2, when the interception component completes the interception of the corresponding medical feature text, the screening unit determines whether the intercepted text is a valid intercepted text according to the character set length of the intercepted medical feature text,
if the character set length of the intercepted medical feature text is smaller than the preset character set length, the screening unit judges that the intercepted text is not an effective intercepted text, and uses the HOOK to remove the intercepted text;
If the character set length of the intercepted medical feature text is larger than the preset character set length, the screening unit judges that the intercepted text is a valid text, and performs part-of-speech tagging on the intercepted text by using jieba segmentation.
With continued reference to fig. 1 and 3, the calculation formula of the complexity evaluation parameter of the grammar of the words is:
wherein S is the complexity evaluation parameter of the grammar of the word, A is the number of misplaced Chinese characters of the grammar components, a is the weight coefficient of the number of misplaced Chinese characters of the grammar components, B is the number of grammar components, B is the weight coefficient of the number of grammar components, C is the average Chinese character number among grammar components, and C is the weight coefficient of the average Chinese character number among grammar components; a=0.4, b=0.2, c=0.4.
With continued reference to fig. 1 and 2, the calculation formula of the patient symptom similarity evaluation parameter or the medical response text similarity evaluation parameter is as follows:
wherein U is a patient symptom similarity evaluation parameter or a medical response text similarity evaluation parameter, E is a cosine similarity between a patient symptom text satisfying a first symptom feature quantity condition and standard symptom sentences stored in the cloud database or a cosine similarity between a medical response text and standard diagnostic sentences stored in the cloud database, E is a cosine similarity weight coefficient, P is an edit distance between a patient symptom text satisfying a first symptom feature quantity condition and standard symptom sentences stored in the cloud database or an edit distance between a medical response text and standard diagnostic sentences stored in the cloud database, P is an edit distance weight coefficient, e+p=1;
When the medical feature text similarity is judged by taking the absolute difference as a standard, the central control module increases the editing distance weight coefficient so that p is more than e;
and when the medical feature text similarity determination is performed by taking the relative difference as a standard, the central control module adjusts the cosine similarity weight coefficient to be higher so that e is more than p.
Specifically, the preferred embodiment of the absolute difference is to determine the content similarity between the amount of the drug in the medical response text and the amount of the drug in the standard diagnosis sentence stored in the cloud database and determine the content similarity between the type of the drug in the medical response text and the type of the drug in the standard diagnosis sentence stored in the cloud database, but not limited to the above embodiment, it will be understood by those skilled in the art that the content similarity determination may be implemented only under the basic framework.
As will be appreciated by those skilled in the art, for the content similarity determination of the relative differences, a preferred embodiment of the relative differences is the content similarity determination of the patient condition severity descriptive text in the medical response text and the condition severity descriptive text in the standard diagnostic statements stored in the cloud database.
Specifically, the cosine similarity is a cosine value of an included angle between two vectors, and the calculation process is not described herein.
According to the system, through the set adjustment process for the weight coefficient, the weight coefficient is adjusted according to different standard conditions, so that the accuracy of similarity detection is improved, and the accuracy of medical inquiry is further improved.
Example 1
In this embodiment 1, when the similarity between the medical diagnosis text output by the AI-model-based medical inquiry system and the standard diagnosis text in the cloud database is detected based on the present invention, the usage amount of the drug is used as the absolute difference, and the weighting unit increases the edit distance weight coefficient so that p > e, in this embodiment 1, p=0.65 times -1 E=0.35, e=cos 30° approximately 0.87, p=5 times, the similarity determination module finds u=0.87×0.35+5×0.65 approximately 3.55.
In the embodiment 1, the medical diagnosis accuracy is improved under the condition of absolute difference by detecting and quantitatively calculating the similarity between the medical diagnosis text and the standard diagnosis text in the cloud database.
Example 2
In this embodiment 2, the standard medical feature recognition granularity is adjusted, the central control module determines three types of adjustment modes for the standard medical feature recognition granularity according to the difference between the medical response text similarity evaluation parameter and the preset response text similarity evaluation parameter under the condition of the preset first similarity evaluation parameter, the difference of the preset first response sentence similarity evaluation parameter is denoted as Δe1, the difference of the preset second response sentence similarity evaluation parameter is denoted as Δe2, the preset first granularity adjustment parameter is denoted as α1, the preset second granularity adjustment parameter is denoted as α2, the preset recognition granularity is denoted as N0, n0=6, wherein Δe1=0.32, Δe2=0.6, α1=2, α2=4,
In this embodiment 2, Δe=0.42 is obtained, the central control module determines Δe1 < Δe2and adjusts the standard medical feature identification granularity to the first identification granularity N 'by using the preset second granularity adjustment coefficient α2, so as to calculate N' =6× (1+4)/2=15.
According to the system, the standard medical feature identification granularity is adjusted to the corresponding value according to the preset first answer sentence similarity, the preset second answer sentence similarity evaluation parameter difference, the preset first granularity adjustment coefficient and the preset second granularity adjustment coefficient, so that the influence on the accuracy of the inquiry due to the inaccuracy of the adjustment of the identification granularity is reduced, and the accuracy of the medical inquiry is improved.
Thus far, the technical solution of the present invention has been described in connection with the preferred embodiments shown in the drawings, but it is easily understood by those skilled in the art that the scope of protection of the present invention is not limited to these specific embodiments. Equivalent modifications and substitutions for related technical features may be made by those skilled in the art without departing from the principles of the present invention, and such modifications and substitutions will be within the scope of the present invention.
Claims (9)
1. An AI model-based medical interrogation system, comprising:
the inquiry module is used for carrying out text dialogue inquiry and comprises a medical inquiry unit used for inquiring medical questions and a response unit connected with the medical inquiry unit and used for responding the medical questions;
the cloud database is connected with the inquiry module and used for storing data updated in real time by the cloud and corresponding inquiry data of a plurality of medical inquiry processes;
the detection module is respectively connected with the inquiry module and the cloud database and comprises a first similarity detection unit which is respectively connected with the medical inquiry unit and the cloud database and used for detecting and calculating the symptom similarity evaluation parameters of the patient, a second similarity detection unit which is respectively connected with the response unit and the cloud database and used for detecting and calculating the similarity evaluation parameters of the medical response text, and a screening unit which is respectively connected with the medical inquiry unit and the response unit and used for screening out the symptom texts of the corresponding patient and the corresponding medical response texts,
the screening unit includes:
an interception component connected with the medical question unit and the response unit respectively, for intercepting patient symptom texts and medical response texts by using a HOOK;
The label generation component is connected with the interception component and is used for carrying out labeling processing and storage on the intercepted medical feature text through jieba segmentation, and the medical feature text comprises a patient symptom text and a medical response text;
wherein the tag generation component comprises:
the paragraph structure serialization plug-in is used for carrying out serialization operation on the medical feature text according to the number of characters between adjacent commas of the medical feature text, and respectively transmitting the serialized medical feature text to a first similarity detection unit and a second similarity detection unit;
the vocabulary diversity serialization plug-in is used for carrying out serialization operation on the medical feature texts according to the number of different vocabularies and the types of the vocabularies, and respectively transmitting the serialized medical feature texts to the first similarity detection unit and the second similarity detection unit;
the grammar complexity plug-in is used for carrying out serialization operation on the medical feature text according to the grammar complexity evaluation parameters of the words and respectively transmitting the serialized medical feature text to the first similarity detection unit and the second similarity detection unit;
the vocabulary structure plug-in is used for carrying out serialization operation on the medical feature text according to whether the vocabulary is an overlapping word or not, and respectively transmitting the serialized medical feature text to the first similarity detection unit and the second similarity detection unit;
A keyword fingerprint plug-in for storing the medical feature text after the removed stop word;
for the paragraph structure serialization plugin, if the number of characters between adjacent commas is an integer multiple of a preset character number unit threshold value which is larger than 0, the paragraph structure serialization plugin takes the number of characters between the current adjacent commas as a vector standard for serialization;
the label generating component is used for arranging and storing punctuation marks according to the bit times when using different serialization plugins to carry out serialization operation on the medical feature text;
the patient symptom similarity evaluation parameters are similarity evaluation parameters of a patient symptom text meeting the first symptom feature quantity condition and standard symptom sentences stored in the cloud database, and the medical response text similarity evaluation parameters are similarity evaluation parameters of a medical response text output by the response unit and standard diagnosis sentences stored in the cloud database;
the central control module is respectively connected with the inquiry module, the cloud database and the detection module and is used for judging whether to call a fixed medical response text in the cloud database according to the patient symptom similarity evaluation parameter detected by the first similarity detection unit and judging whether the accuracy of the medical response text is in an allowable range according to the medical response text similarity evaluation parameter detected by the second similarity detection unit and adjusting the standard medical feature recognition granularity to a corresponding granularity;
The first symptom characteristic quantity condition is that the symptom characteristic quantity in the patient symptom text output by the medical questioning unit is larger than the preset symptom characteristic quantity.
2. The AI model-based medical consultation system of claim 1, wherein the central control module determines whether to call a fixed medical response text in a cloud database based on the patient symptom similarity evaluation parameter detected by the first similarity detection unit,
if the patient symptom similarity evaluation parameter is smaller than the preset symptom similarity evaluation parameter, the central control module judges that the complexity of the patient symptom text output by the medical questioning unit is beyond the allowable range, and the central control module retrieves the fixed medical response text in the cloud database and sends the fixed medical response text to the response unit after the retrieval is completed;
and if the patient symptom similarity evaluation parameter is greater than or equal to the preset symptom similarity evaluation parameter, the central control module judges that the complexity of the patient symptom text is within the allowable range and controls the response unit to output the corresponding medical response text.
3. The AI model-based medical consultation system according to claim 2, wherein the central control module determines whether the accuracy of the medical response is within two types of decision modes within an allowable range according to the calculated medical response text similarity evaluation parameter detected by the second similarity detection unit, wherein,
The first type of judgment mode is that the central control module judges that the accuracy of the medical response is lower than an allowable range under the condition of presetting a first similarity evaluation parameter, and adjusts the standard medical feature identification granularity to a corresponding granularity by calculating the difference value between the similarity evaluation parameter of the medical response text and the similarity evaluation parameter of the preset response text;
the second type of judgment mode is that the central control module judges that the accuracy of the medical response is within an allowable range under the condition of presetting a second similarity evaluation parameter;
the condition of the preset first similarity evaluation parameter is that the medical response text similarity evaluation parameter is smaller than or equal to the preset response text similarity evaluation parameter; the condition of the preset second similarity evaluation parameter is that the medical response text similarity evaluation parameter is larger than the preset response text similarity evaluation parameter.
4. The AI model-based medical consultation system of claim 3 wherein the central control module determines three types of adjustment for standard medical feature recognition granularity based on differences between the medical response text similarity evaluation parameter and the preset response text similarity evaluation parameter under the preset first similarity evaluation parameter conditions,
The first type of adjustment mode is that the central control module adjusts the standard medical feature identification granularity to a preset identification granularity under the condition of presetting a first similarity evaluation parameter difference value;
the second type of adjustment mode is that the central control module adjusts the standard medical feature identification granularity to the first identification granularity by using a preset second granularity adjustment coefficient under the condition of presetting a second similarity evaluation parameter difference value;
the third type of adjustment mode is that the central control module adjusts the standard medical characteristic identification granularity to the second identification granularity by using a preset first granularity adjustment coefficient under the condition of presetting a third similarity evaluation parameter difference value;
the difference value condition of the preset first similarity evaluation parameter is that the difference value of the similarity evaluation parameter of the medical response text and the similarity evaluation parameter of the preset response text is smaller than or equal to the difference value of the similarity evaluation parameter of the preset first response sentence; the difference value condition of the preset second similarity evaluation parameter is that the difference value of the medical response text similarity evaluation parameter and the preset response text similarity evaluation parameter is larger than the difference value of the preset first response sentence similarity evaluation parameter and smaller than or equal to the difference value of the preset second response sentence similarity evaluation parameter; the difference value condition of the preset third similarity evaluation parameter is that the difference value between the medical response text similarity evaluation parameter and the preset response text similarity evaluation parameter is larger than the difference value of the preset second response sentence similarity evaluation parameter; the preset first answer sentence similarity evaluation parameter difference value is smaller than the preset second answer sentence similarity evaluation parameter difference value, and the preset first granularity adjustment coefficient is smaller than the preset second granularity adjustment coefficient.
5. The AI model-based medical interrogation system of claim 4, wherein the standard medical signature recognition granularity is granularity for disease symptom signatures in re-questioning notifications issued by the system to an interrogating patient.
6. The AI model-based medical interrogation system of claim 1, wherein the interception component determines whether to intercept a medical feature text based on a number of repeated words in the medical feature text,
if the number of repeated words in the medical feature text is greater than or equal to the standard number of words in the medical feature text, the interception component judges that the medical feature text is intercepted;
the calculation formula of the repeated word number ratio in the medical feature text is as follows:
wherein G is the number of repeated words in the medical feature text to be occupied, M is the total number of repeated words in the medical feature text, and F is the total number of words in the medical feature text.
7. The AI model-based medical consultation system of claim 6, wherein the screening unit determines whether the intercepted text is a valid intercepted text based on a character set length of the intercepted medical feature text when the intercepting component completes the interception of the corresponding medical feature text,
If the character set length of the intercepted medical feature text is smaller than the preset character set length, the screening unit judges that the intercepted text is not an effective intercepted text, and uses the HOOK to remove the intercepted text;
if the character set length of the intercepted medical feature text is larger than the preset character set length, the screening unit judges that the intercepted text is a valid text, and performs part-of-speech tagging on the intercepted text by using jieba segmentation.
8. The AI model-based medical interrogation system of claim 7, wherein the computational formula for the complexity evaluation parameter of the grammar of words is:
wherein S is the complexity evaluation parameter of the grammar of the word, A is the number of misplaced Chinese characters of the grammar components, a is the weight coefficient of the number of misplaced Chinese characters of the grammar components, B is the number of grammar components, B is the weight coefficient of the number of grammar components, C is the average Chinese character number among grammar components, and C is the weight coefficient of the average Chinese character number among grammar components; a=0.4, b=0.2, c=0.4.
9. The AI model-based medical interrogation system of claim 8, wherein the patient symptom similarity evaluation parameter or medical response text similarity evaluation parameter is calculated as:
Wherein U is a patient symptom similarity evaluation parameter or a medical response text similarity evaluation parameter, E is a cosine similarity between a patient symptom text satisfying a first symptom feature quantity condition and standard symptom sentences stored in the cloud database or a cosine similarity between a medical response text and standard diagnostic sentences stored in the cloud database, E is a cosine similarity weight coefficient, P is an edit distance between a patient symptom text satisfying a first symptom feature quantity condition and standard symptom sentences stored in the cloud database or an edit distance between a medical response text and standard diagnostic sentences stored in the cloud database, P is an edit distance weight coefficient, e+p=1;
when the medical feature text similarity is judged by taking the absolute difference as a standard, the central control module increases the editing distance weight coefficient so that p is more than e;
and when the medical feature text similarity determination is performed by taking the relative difference as a standard, the central control module adjusts the cosine similarity weight coefficient to be higher so that e is more than p.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310530384.7A CN116313162B (en) | 2023-05-12 | 2023-05-12 | Medical inquiry system based on AI model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310530384.7A CN116313162B (en) | 2023-05-12 | 2023-05-12 | Medical inquiry system based on AI model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116313162A CN116313162A (en) | 2023-06-23 |
| CN116313162B true CN116313162B (en) | 2023-08-18 |
Family
ID=86789021
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310530384.7A Active CN116313162B (en) | 2023-05-12 | 2023-05-12 | Medical inquiry system based on AI model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116313162B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117352195A (en) * | 2023-10-16 | 2024-01-05 | 广州市诺祥康医药科技有限公司 | Online inquiry method and system based on Internet of things |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109545391A (en) * | 2018-10-26 | 2019-03-29 | 曾警卫 | A kind of intelligent medical automatically request-answering system and application method based on deep learning |
| WO2020006835A1 (en) * | 2018-07-03 | 2020-01-09 | 平安科技（深圳）有限公司 | Customer service method, apparatus, and device for engaging in multiple rounds of question and answer, and storage medium |
| CN111564210A (en) * | 2020-04-01 | 2020-08-21 | 心医国际数字医疗系统(大连)有限公司 | Intelligent diagnosis guiding method and device, electronic equipment and storage medium |
| CN112632257A (en) * | 2020-12-29 | 2021-04-09 | 深圳赛安特技术服务有限公司 | Question processing method and device based on semantic matching, terminal and storage medium |
| KR102261092B1 (en) * | 2021-01-20 | 2021-06-03 | 이창엽 | Medical diagnosis management system |
| CN113707307A (en) * | 2021-08-30 | 2021-11-26 | 康键信息技术（深圳）有限公司 | Disease analysis method and device, electronic equipment and storage medium |
| CN113921149A (en) * | 2021-10-26 | 2022-01-11 | 杭州祺鲸科技有限公司 | Standard symptom extraction method based on multiple rounds of inquiry |
| CN115658858A (en) * | 2022-09-28 | 2023-01-31 | 平安科技（深圳）有限公司 | Dialog recommendation method based on artificial intelligence and related equipment |
| WO2023035623A1 (en) * | 2021-09-09 | 2023-03-16 | 平安科技（深圳）有限公司 | Answer corpus generation method based on artificial intelligence, and related device |
-
2023
- 2023-05-12 CN CN202310530384.7A patent/CN116313162B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020006835A1 (en) * | 2018-07-03 | 2020-01-09 | 平安科技（深圳）有限公司 | Customer service method, apparatus, and device for engaging in multiple rounds of question and answer, and storage medium |
| CN109545391A (en) * | 2018-10-26 | 2019-03-29 | 曾警卫 | A kind of intelligent medical automatically request-answering system and application method based on deep learning |
| CN111564210A (en) * | 2020-04-01 | 2020-08-21 | 心医国际数字医疗系统(大连)有限公司 | Intelligent diagnosis guiding method and device, electronic equipment and storage medium |
| CN112632257A (en) * | 2020-12-29 | 2021-04-09 | 深圳赛安特技术服务有限公司 | Question processing method and device based on semantic matching, terminal and storage medium |
| KR102261092B1 (en) * | 2021-01-20 | 2021-06-03 | 이창엽 | Medical diagnosis management system |
| CN113707307A (en) * | 2021-08-30 | 2021-11-26 | 康键信息技术（深圳）有限公司 | Disease analysis method and device, electronic equipment and storage medium |
| WO2023035623A1 (en) * | 2021-09-09 | 2023-03-16 | 平安科技（深圳）有限公司 | Answer corpus generation method based on artificial intelligence, and related device |
| CN113921149A (en) * | 2021-10-26 | 2022-01-11 | 杭州祺鲸科技有限公司 | Standard symptom extraction method based on multiple rounds of inquiry |
| CN115658858A (en) * | 2022-09-28 | 2023-01-31 | 平安科技（深圳）有限公司 | Dialog recommendation method based on artificial intelligence and related equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116313162A (en) | 2023-06-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116313162B (en) | Medical inquiry system based on AI model | |
| CN113241138B (en) | Medical event information extraction method and device, computer equipment and storage medium | |
| CN113270197A (en) | Health prediction method, system and storage medium based on artificial intelligence | |
| CN116153505B (en) | Intelligent critical patient sign identification method and system based on medical pressure sensor | |
| CN110472209B (en) | Deep learning-based table generation method and device and computer equipment | |
| CN113822366A (en) | Service index abnormality detection method and device, electronic equipment and storage medium | |
| CN115238828A (en) | Chromatograph fault monitoring method and device | |
| CN112734202A (en) | Medical capability evaluation method, device, equipment and medium based on electronic medical record | |
| CN113438114A (en) | Method, device, equipment and storage medium for monitoring running state of Internet system | |
| US20200104412A1 (en) | Combined data driven and knowledge driven analytics | |
| WO2023207557A1 (en) | Method and apparatus for evaluating robustness of service prediction model, and computing device | |
| US10224036B2 (en) | Automated identification of verbal records using boosted classifiers to improve a textual transcript | |
| CN111161880B (en) | Medical information classification method and device based on classification model and computer equipment | |
| CN110457450A (en) | Answer generation method and relevant device based on neural network model | |
| CN113593692A (en) | Data processing method based on big data intelligent medical treatment and cloud computing server | |
| CN113516454B (en) | Enterprise information management system and method based on cloud computing platform | |
| CN114936600A (en) | Document abnormity monitoring method, device, equipment and storage medium | |
| CN114420307A (en) | Artificial intelligence-based public health event registration method and device and electronic equipment | |
| CN116487049B (en) | Unplanned reentry event processing system based on decision analysis | |
| CN112163585A (en) | Text auditing method and device, computer equipment and storage medium | |
| US10102860B2 (en) | Common phrase identification and language dictation recognition systems and methods for using the same | |
| US20230298765A1 (en) | Computer-implemented process for predicting health risk in real-time | |
| KR102388184B1 (en) | Collection, classification and management system of public information for detecting and controlling undeclared nuclear fuel cycle-related activities and method thereof | |
| CN117807105A (en) | Analog quantity calibration method and related device | |
| CN116384370A (en) | Big data security analysis method and system for online service session interaction |
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 |