CN106776606A - Retrieval device and search method based on electronic health record database - Google Patents
Retrieval device and search method based on electronic health record database Download PDFInfo
- Publication number
- CN106776606A CN106776606A CN201510809140.8A CN201510809140A CN106776606A CN 106776606 A CN106776606 A CN 106776606A CN 201510809140 A CN201510809140 A CN 201510809140A CN 106776606 A CN106776606 A CN 106776606A
- Authority
- CN
- China
- Prior art keywords
- keyword
- electronic health
- health record
- correlation
- degree
- 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.)
- Pending
Links
Classifications
-
- 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
- G06F16/3349—Reuse of stored results of previous queries
-
- 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
- G06F16/3346—Query execution using probabilistic model
-
- G06Q50/24—
Abstract
Retrieval device and search method the present invention relates to be based on electronic health record database, retrieval device include:Input unit, the first keyword for being input into user's determination;First search part, it is based on electronic health record database described in the first key search, obtains the first retrieval result;Whether judging part, the number that it judges the first retrieval result is 0, if 0, then returns to prompt message, according to the input of user if being not 0, judges whether the first retrieval result is qualified;Calculating part, when judging part judges that the first retrieval result is unqualified, calculates the degree of correlation of the keyword in addition to first keyword and the first keyword in multiple electronic health records, recommended unit, and it recommends the keyword of degree of correlation maximum as the second keyword;Second search part, it is based on the second key search electronic health record database, obtains the second retrieval result;Display part, it can show to prompt message, the first retrieval result and the second retrieval result.
Description
Technical field
The present invention relates to a kind of retrieval device and search method based on electronic health record database, more particularly to pushed away according to the degree of correlation
Recommend the retrieval device and search method of keyword.
Background technology
Case history searching system based on electronic health record database is widely used gradually in medical profession, makes personnel used in connection with special
It is medical history information that medical personnel can quickly obtain close patient, to improve the accuracy to medical diagnosis on disease, while being easy to
Clinical research worker meets its scientific research demand more accurately to the inquiry and extraction of required case history.For electronic health record number
Arrangement extraction, electronic health record number can be carried out by Chinese word segmentation method of the prior art according to the keyword in the electronic health record in storehouse
It is by the electronic health record after Chinese word segmentation treatment according to electronic health record included in storehouse.
One of important core function of case history searching system is the retrieval of case history, and similar system is true according to user on the market at present
Fixed keyword is retrieved, and shows retrieval result.But needed for sometimes retrieval result is not necessarily user, cause
Retrieval effectiveness is unsatisfactory.
The searching system recommended the keyword of electronic health record is not directed in the prior art.
The content of the invention
The present invention provides a kind of retrieval apparatus and method, in the case where keyword determined by user is undesirable, according to
The degree of correlation carries out keyword recommendation, to reach the purpose of more accurate retrieval result.
An aspect of of the present present invention provides a kind of retrieval device based on electronic health record database, and the electronic health record database includes
Through multiple electronic health records of Chinese word segmentation, the retrieval device includes:Input unit, the first keyword for being input into user's determination;
First search part, its be based on first key search described in electronic health record database, obtain the first retrieval result;Judge
Whether portion, the number that it judges first retrieval result is 0, if 0, then returns to prompt message, if not being 0
Then according to the input of user, judge whether first retrieval result is qualified;Calculating part, when the judging part judges described
When one retrieval result is unqualified, the keyword two-dimensional matrix of the multiple electronic health record is set up, and based on the keyword Two-Dimensional Moment
The keyword in addition to first keyword that battle array is calculated in the multiple electronic health record is related to first keyword
Degree, it with the keyword in the multiple electronic health record is row name that the keyword two-dimensional matrix is, with the multiple electronic health record
Medical record number be to go the binaryzation matrix of name, wherein when the keyword for thering is row name to be enumerated in electronic health record, by it at this
The value of the place ranks in keyword two-dimensional matrix is designated as 1, is otherwise designated as 0;Recommended unit, it is calculated according to the calculating part
The keyword in addition to first keyword in the multiple electronic health record for obtaining is related to first keyword
Degree, recommends the maximum keyword of the degree of correlation as the second keyword;Second search part, it is based on second key search
The electronic health record database, obtains the second retrieval result;Display part, it is to the prompt message, the first retrieval knot
Fruit and second retrieval result are shown.
According to above-mentioned retrieval device, user can in the case where with keyword determined by oneself desired result cannot be retrieved,
Further retrieved according to the keyword that device is recommended.
Further, the judging part can interpolate that whether the number of first keyword is single;When the judging part is sentenced
Break first keyword for it is single when, calculating part be based on following formula calculate in the multiple electronic health record except described first
The degree of correlation R of keyword outside keyword and first keywordkf,
Formula 1
Wherein, p represents that electronic health record numbers be expert at sequence number, p=1,2,3...., j;J represents the electronics in electronic medical record system
The total number of case history, j is positive integer;K represent the first search key where row sequence number;F is represented except the first keyword
The sequence number of keyword column in addition, f=1 ... k-1, k+1 ... i;I is positive integer, as all keys in electronic medical recordses system
The total number of word (without repeating);Z is the value corresponding to the corresponding ranks in the keyword two-dimensional matrix.
According to above-mentioned retrieval device, using the teaching of the invention it is possible to provide user determine the first keyword be it is single in the case of a kind of degree of correlation
The mode of calculating.
Further, judging part can interpolate that whether the number of first keyword is single;When the judging part judges
When one keyword is two or more word, the calculating part be based on following formula calculate in the multiple electronic health record except described
The degree of correlation Ln of keyword outside the first keyword and first keyword,
Formula 2
Wherein, p represents that electronic health record numbers be expert at sequence number, p=1,2,3...., j;J represents the electronics in electronic medical record system
The total number of case history, j is positive integer;A, b, c...t be the first keyword each where row sequence number, n is except the first keyword
The sequence number of keyword column in addition.
According to above-mentioned retrieval device, using the teaching of the invention it is possible to provide a kind of phase in the case where the first keyword that user determines is for two or more
The mode that Guan Du is calculated.
Further, when the key in addition to first keyword in the multiple electronic health record that calculating part is calculated
There is same value in the degree of correlation of word and first keyword and when being all maximum value, also phase can respectively be calculated based on following formula
Guan Du is the weights We with each keyword column of value,
Formula 3
Wherein, p represents that electronic health record numbers be expert at sequence number, p=1,2,3...., j;J represents the electronics disease in electronic medical record system
The total number gone through, j is positive integer;E represents that the degree of correlation produces the sequence number with the keyword column of value,
It is the second keyword that the recommended unit recommends the maximum keywords of weights We, if the degree of correlation is with each keyword institute of value
Also it is with value, then to select the keyword that any one degree of correlation is maximum value as the second keyword in the weights We of row.
According to above-mentioned retrieval device, using the teaching of the invention it is possible to provide in the keyword in addition to first keyword and first keyword
There is same value in the degree of correlation and when being all maximum value, select a kind of mode of the second keyword.
Another aspect of the present invention provides a kind of retrieval mode based on electronic health record database, the electronic health record database bag
Include through multiple electronic health records of Chinese word segmentation, the search method includes:Input step, the first pass for being input into user's determination
Key word;First searching step, its be based on first key search described in electronic health record database, obtain the first retrieval knot
Really;Judge step, whether the number that it judges first retrieval result is 0, if 0, then returns to prompt message,
According to the input of user if being not 0, judge whether first retrieval result is qualified;Calculation procedure, when the judgement
When step judges that first retrieval result is unqualified, the keyword two-dimensional matrix of the multiple electronic health record is set up, and be based on
The keyword two-dimensional matrix calculates keyword in addition to first keyword in the multiple electronic health record with described the
The degree of correlation of one keyword, it with the keyword in the multiple electronic health record is row name that the keyword two-dimensional matrix is, with institute
The medical record number for stating multiple electronic health records is to go the binaryzation matrix of name, wherein when the key for having row name to be enumerated in electronic health record
During word, the value of its place ranks in the keyword two-dimensional matrix is designated as 1, is otherwise designated as 0;Recommendation step, its root
According to the keyword in the multiple electronic health record that the calculation procedure is calculated in addition to first keyword and institute
The degree of correlation of the first keyword is stated, recommends the maximum keyword of the degree of correlation as the second keyword;Second searching step, its base
In electronic health record database described in second key search, the second retrieval result is obtained;Step display, it is carried to described
Show that information, first retrieval result and second retrieval result are shown.
Brief description of the drawings
Fig. 1 is the functional block diagram of the retrieval device for showing an embodiment of the present invention.
Fig. 2 is an example of keyword two-dimensional matrix.
Fig. 3 shows the flow chart of the search method of one embodiment of the invention.
Specific embodiment
Hereinafter, referring to the drawings, preferred embodiment it is described in detail to of the invention.Here, in the explanation of accompanying drawing
In, identical symbol is marked to key element identically or comparably, the repetitive description thereof will be omitted.
Fig. 1 is the functional block diagram of the retrieval device for showing an embodiment of the present invention.As shown in figure 1, retrieval device 100 is wrapped
Include, input unit 10, the first search part 20, judging part 30, calculating part 40, recommended unit 50, the second search part 60 and
Display part 70.
Retrieval device 100 is the retrieval device that suitable electronic health record is carried out based on electronic health record database (not shown),
The multiple electronics by the keyword extraction in each electronic health record out through Chinese word segmentation are included in electronic health record database
Case history, retrieval device can be connected by communication network etc. with electronic health record database.The coverage of the electronic health record database
Can be the electronic health record of the patient of hospital of the whole city, even the whole nation, the networking data storehouse of global range.
Input unit 10 is used to be input into the first keyword of user's determination.User is typically doctor, researcher etc., and it can basis
It needs to be determined that the first keyword, such as in diagnosis during patient with " cough " symptom, its may using " cough " as
First keyword come carry out retrieve similar patient case history think after diagnosis make reference.
First search part 20 retrieves electronic health record database based on the first keyword being input into by input unit 10, obtains
One retrieval result.First retrieval result refers to the electronic health record obtained according to the first key search.
Judging part 30 first judges whether the number of first retrieval result is 0 and (does not detect matching first keyword
Case history), if 0, then prompt message is returned to, the prompt message can be shown in display part described later 70, be used to remind use
Family redefines the first new keyword;If the number of the first retrieval result is not 0, judging part 30 is according to the straight of user
Input or the input of the user sent from input unit 10 is connect to judge to be examined according to the first keyword by the first search part 20
Whether rope obtains the first retrieval result qualified.Whether user meets expected feedback result, Ke Yitong for the first retrieval result
Cross after input unit 10 is input into and judging part 30 is fed back to by input unit 10, can also directly input to judging part 30.
Calculating part 40 can calculate the degree of correlation of the keyword and the first keyword in addition to the first keyword.When judging part 30
When judging that the first retrieval result is unqualified, the keyword of multiple electronic health records that calculating part 40 is set up in electronic health record database
Two-dimensional matrix, and the key in addition to first keyword in multiple electronic health records is calculated based on the keyword two-dimensional matrix
The degree of correlation of word and first keyword.As shown in Fig. 2 above-mentioned keyword two-dimensional matrix is with electronic health record database
Multiple electronic health records in mutually unduplicated keyword be row name (such as " palpitaition ", " expiratory dyspnea ", " cough ", " uncomfortable in chest ",
" oedema " etc.), and with the medical record number of the multiple electronic health record be row name (for example, ID1, ID2, ID3, ID4 etc.)
Binaryzation matrix.When the row name for having the keyword two-dimensional matrix in certain electronic health record in electronic health record database is enumerated
Keyword when, the value of the place ranks by the electronic health record in keyword two-dimensional matrix is designated as 1, is otherwise designated as 0.
Each in addition to first keyword in multiple electronic health records that recommended unit 50 is calculated according to calculating part 40
Keyword and the degree of correlation of first keyword, recommend the wherein maximum keyword of the degree of correlation as the second keyword.
The second key search electronic health record database that second search part 60 is recommended based on recommended unit 50, obtains the second inspection
Hitch is really.
Display part 70 can be to prompt message, or the first retrieval result or the second retrieval result are shown.
The calculating part 40 of calculating on to(for) the keyword in addition to first keyword and the degree of correlation of first keyword,
That is, the calculating of the degree of correlation in matrix between each row, those skilled in the art can be according to it needs to be determined that suitable relatedness computation
Mode.Below, the calculated examples of the degree of correlation under some different situations are given.
Hereinafter, illustrate how to carry out the retrieval based on the first keyword and the inspection when the first keyword by specific calculated example
Retrieval based on the second keyword when hitch fruit is unqualified.
Embodiment 1
First, in embodiment 1, input unit 10 is transfused to the first keyword " cough " determined by user.First retrieval
Portion 20 is retrieved based on first keyword " cough " to electronic health record database, obtains the first retrieval result, also
It is to say to retrieve all case histories related to first keyword " cough ", and is shown in display part 70, can be examined in this example
It is ID1, three case histories of ID3, ID4 that rope is numbered to case in electronic medical recordses database.Thus judging part 30 judges first
The number of retrieval result is not 0.Now, may be dissatisfied to first retrieval result using the user of retrieval device,
For example, it is believed that the electronic health record amount that its retrieval is obtained is too huge, or thinks to retrieve the reference value of the electronic health record for obtaining
It is not high, and the underproof feedback of the first retrieval result is made, such feedback for the first retrieval result can be by input
Portion 10 is sent to judging part 30, can also be transmitted directly to judging part 30.Judging part 30 be based on user input for first
The feedback of retrieval result, judges whether the first retrieval result is qualified.If it is determined that the first retrieval result is qualified, then in display part
70 the first retrieval results of display;If it is determined that the first retrieval result is unqualified, the follow-up degree of correlation is carried out by calculating part 40
Calculate.
In embodiment 1, whether judging part 30 can also be single to judge to the first keyword.With the present embodiment 1
As a example by, judging part 30 judges the first keyword " cough " for single, then calculating part 40 is first with electronic health record database
Multiple electronic health records in mutually unduplicated keyword be row name, and with the medical record number of the multiple electronic health record be row name structure
Building two-dimensional matrix, as follows (example of the two-dimensional matrix is that for the purpose of simplifying the description, the two-dimensional matrix actually constructed can be with than it
It is huge a lot), when having the keyword that the row name of the keyword two-dimensional matrix is enumerated in specific electronic health record, by the electronics
The value of place ranks of the case history in keyword two-dimensional matrix is designated as 1, is otherwise designated as 0, as follows:
It is single keyword " cough " by the first keyword determined by the user that input unit is input into this example, now,
It is other keywords " stomachache ", " fever ", " nasal obstruction " and the first keyword " cough " that calculating part 40 is calculative
The degree of correlation is respectively how many.
After electronic health record two-dimensional matrix is constructed, calculating part 40 be based on following formula calculate in the multiple electronic health record except institute
State the degree of correlation R of the keyword and first keyword outside the first keywordkf。
Formula 1
Wherein, p represents that electronic health record numbers be expert at sequence number, p=1,2,3...., j;J represents the electronics in electronic medical record system
The total number of case history, j is positive integer;K represent the first search key where row sequence number;F is represented except the first keyword
The sequence number of keyword column in addition, f=1 ... k-1, k+1 ... i;I is positive integer, all keys as in electronic medical record system
The total number of word (without repeating);Z is the value corresponding to the corresponding ranks in the keyword two-dimensional matrix.
According to the above, for example, R13What is represented is first keyword of the keyword with the serial number 1 of row of the serial number 3 of row
The degree of correlation.
According to formula 1,
Thus, the keyword in addition to the first keyword and first that recommended unit 50 can be calculated according to above-mentioned calculating part are crucial
The degree of correlation of word recommends wherein degree of correlation highest.By taking embodiment 1 as an example, wherein, the keyword of the serial number 4 of row
The degree of correlation R of " nasal obstruction " and the first keyword " cough "14Be highest, the keyword " fever " of the serial number 3 of row with
The degree of correlation R of the first keyword " cough "13For minimum.Therefore, it is recommended that " nasal obstruction " is recommended as the second keyword by portion 50.
Second search part 60 is retrieved according to second keyword " nasal obstruction " to electronic health record database, is obtained further
Retrieval result, i.e. the second retrieval result, and the second retrieval result is shown in display part 70.
Separately, also there are the situation that the first keyword is two or more (including two), the example that for example embodiment 2 is enumerated.
Embodiment 2
In example 2, input unit 10 is transfused to the first keyword " cough " and " fever " determined by user.The
One search part 20 is retrieved based on the first keyword " cough " and " fever " to electronic health record database, obtains first
Retrieval result, that is to say, that retrieve all case histories related to the first keyword " cough " and " fever ", and show
In display part 70.The case history that case numbering in electronic medical recordses database is ID1 can be retrieved in this example.Judging part 30 by
This judges that the number of the first retrieval result is not 0.Now, may be to first retrieval result using the user of retrieval device
It is dissatisfied, and the underproof feedback of the first retrieval result is made, such feedback for the first retrieval result can be by defeated
Enter portion 10 and be sent to judging part 30, can also be transmitted directly to judging part 30.Judging part 30 be based on user input for the
The feedback of one retrieval result, judges whether the first retrieval result is qualified.If it is determined that the first retrieval result is qualified, then in display
Portion 70 shows the first retrieval result;If it is determined that the first retrieval result is unqualified, the follow-up degree of correlation is carried out by calculating part 40
Calculating.
In example 2, whether judging part 30 can be equally single to judge to the first keyword.With embodiment 2
As a example by, judging part 30 judges the first keyword " cough " and " fever " is more than two words.Now, calculating part
40 is row name first with mutual unduplicated keyword in the multiple electronic health records in electronic health record database, and with the multiple electricity
To construct two-dimensional matrix as follows for row name for the medical record number of sub- case history, there is the keyword two-dimensional matrix in specific electronic health record
Row name enumerated keyword when, the value of the place ranks by the electronic health record in keyword two-dimensional matrix is designated as 1, no
Then it is designated as 0, it is as follows:
In this example, by input unit be input into user determined by the first keyword be keyword " cough " and " fever ",
Now, calculating part 40 it is calculative be other keywords " stomachache ", " nasal obstruction " and the first keyword " cough " and
The degree of correlation of " fever " is respectively how many.
After electronic health record two-dimensional matrix is constructed, calculating part 40 calculates the multiple electronic health record based on formula 2 as shown below
In keyword in addition to first keyword and the first keyword degree of correlation Ln.
Formula 2
Wherein, p represents that electronic health record numbers be expert at sequence number, p=1,2,3...., j;J represents the electronics in electronic medical record system
The total number of case history, j is positive integer;A, b, c...t be the first keyword each where row sequence number, n is except the first keyword
The sequence number of keyword column in addition.
According to the above, for example, L2What is represented is the degree of correlation of the keyword " stomachache " with the first keyword of the serial number 2 of row,
L4Represent the keyword " nasal obstruction " of the serial number 4 of row and the degree of correlation of the first keyword.
According to formula 2, L2=0+0+0+1=1
L4=2+1+1+0=4
Thus, the keyword and first in addition to the first keyword that recommended unit 50 can be calculated according to above-mentioned calculating part is closed
The degree of correlation of key word recommends wherein degree of correlation highest.By taking embodiment 2 as an example, wherein, the key of the serial number 4 of row
Word " nasal obstruction " and the first keyword " cough " and the degree of correlation L of " fever "4It is highest, the keyword of the serial number 2 of row
" stomachache " and the first keyword " cough " and the degree of correlation L of " fever "2For minimum, therefore, it is recommended that " nose is recommended in portion 50
Thiophene " is used as the second keyword.
Second search part 60 is retrieved according to second keyword " nasal obstruction " to electronic health record database, is obtained further
Retrieval result, i.e. the second retrieval result, and the second retrieval result is shown in display part 70.
Although above-described embodiment 1 gives with embodiment 2 and calculates the first key respectively in two kinds of different relatedness computation modes
The example that word is single and the first keyword is more than two situations, but those skilled in the art are also dependent on concrete condition
Calculated by a kind of general calculation the first keyword be single or two or more in the case of the degree of correlation.
Embodiment 3
In addition, when the key in addition to first keyword in the multiple electronic health record that the calculating part is calculated
When having same value in the degree of correlation of word and first keyword and being all maximum, then phase can respectively be calculated based on following formula
Guan Du is the weights We with each keyword column of value,
Formula 3
Wherein, p represents that electronic health record numbers be expert at sequence number, p=1,2,3...., j;J represents the electronics disease in electronic medical record system
The total number gone through, j is positive integer;E represents that the degree of correlation produces the sequence number with the keyword column of value,
It is the second keyword that recommended unit 50 recommends the maximum keywords of weights We.
If according to above-mentioned formula 3, the weights We for being calculated each keyword column that the degree of correlation is same value is also same value,
The keyword that any one degree of correlation is maximum value is then selected as the second keyword.
It is assumed that in embodiment 3, input unit 10 is transfused to the first keyword " cough " and " expectoration " determined by user.
First search part 20 is retrieved based on the first keyword " cough " and " expectoration " to electronic health record database.Judging part
30 judge that the number of the first retrieval result is not 0.And judging part 30 is based on the anti-for the first retrieval result of user input
Feedback, judges whether the first retrieval result is qualified.Assuming that judging the first retrieval result for unqualified in this example.
Next, it is determined that portion 30 judges the first keyword " cough " and " expectoration " is more than two words, calculating part 40
It is determined that calculating the degree of correlation of the keyword and the first keyword in addition to the first keyword with formula 2.Calculating part 40 first with
Mutual unduplicated keyword is row name in multiple electronic health records in electronic health record database, and with the multiple electronic health record
Two-dimensional matrix is as follows for row name is constructed for medical record number, the row name institute for having the keyword two-dimensional matrix in specific electronic health record
During the keyword for enumerating, the value of the place ranks by the electronic health record in keyword two-dimensional matrix is designated as 1, is otherwise designated as 0,
It is as follows:
Obtained according to formula 2, the keyword " uncomfortable in chest " of the serial number 2 of row and the degree of correlation L of the first keyword2, row sequence number
It is 3 keyword " asthma " and the degree of correlation L of the first keyword3, and row serial number 5 keyword " edema of lower extremity "
L5Concrete numerical value.
L2=3
L3=3
L5=2
Understand, degree of correlation magnitude relationship is, L2=L3>L5.Occur in that the keyword in addition to first keyword and institute
State the situation for having same value in the degree of correlation of the first keyword.
Now, these weighted values with the value degree of correlation can be calculated according to formula below 3, to select the wherein big work of weight
It is the second keyword.
Formula 3
Wherein, p represents that electronic health record numbers be expert at sequence number, p=1,2,3...., j;J represents the electronics disease in electronic medical record system
The total number gone through, j is positive integer;E represents that the degree of correlation produces the sequence number with the keyword column of value.
According to formula 3, there is the L with value2And L3Respective weights it is different,
W2=1+1+0+0+0=2
W3=1+1+0+0+1=3.
Therefore, in embodiment 3, the value of final choice weight is the keyword " asthma " of maximum row serial number 3 as the
Two keywords are retrieved, and the second retrieval result is shown in into display part.
Fig. 3 shows the flow chart of the search method of one embodiment of the invention.As shown in figure 3, in step s 11, input is used
The first keyword that family determines, for example, " cough ".In step s 12, according to the first key search electronic health record number
According to storehouse, the first retrieval result is obtained.Then, in step s 13, whether the number for judging the first retrieval result is 0, such as
Fruit judged result is yes, then send prompt message and the display reminding information in step S18, to point out user to redefine
First keyword;If the judged result in step S13 is no, step S14 is entered into.In step S14, according to
The input at family, judges whether the first retrieval result is qualified, if it is judged that being yes, then enters directly into step S18 and shows
First retrieval result;If the judged result of step S14 is no, step S15 is entered into.In step S15, set up many
The keyword two-dimensional matrix of individual electronic health record, calculates the degree of correlation of the keyword and the first keyword in addition to the first keyword,
Calculating on the degree of correlation, those skilled in the art can as needed determine specific relatedness computation mode, also can be as above
Literary embodiment 1 and embodiment 2 carry out the calculating of the degree of correlation like that.In step S16, according to removing that step S15 is calculated
Keyword outside first keyword with the first keyword degree of correlation, recommend the wherein maximum keyword of the degree of correlation as second
Keyword.In step S17, based on the second key search electronic health record database that step S16 is recommended, second is obtained
Retrieval result, then, enters into step S18, shows second retrieval result.
Furthermore it is possible to when step S14 judges the first retrieval result for unqualified (no), determine whether the first keyword
It is single or two or more, and selects different relatedness computation methods accordingly, such as by the meter of above-mentioned formula 1
Calculation method or the computational methods by formula 2.
In addition, in step S15, if the phase of the keyword in addition to the first keyword being calculated and the first keyword
Guan Du occurs with value and is maximum value, then the pass occurred with the value degree of correlation can be further calculated according to above-mentioned formula 3
The respective weight of key word, is maximum keyword as the second keyword using the value of recommending weight.In addition, if the degree of correlation
The weights We for being each keyword column of same value is also that with value, then it is maximum value that can select any one degree of correlation
Keyword is used as the second keyword.
Embodiments of the present invention are illustrated above, but these implementation methods are illustrative only, and without limit
Determine the intention of invention scope.These implementation methods can be implemented by other various forms, in the scope without departing from inventive concept
Inside carry out various omissions, displacement, change, combination.These implementation methods and its deformation are included in invention scope and master
While in purport, in the invention being also contained in described in claims and the scope impartial with it.
Claims (8)
1. a kind of retrieval device based on electronic health record database, the electronic health record database is included through the multiple electricity of Chinese word segmentation
Sub- case history, it is characterised in that the retrieval device includes:
Input unit, the first keyword for being input into user's determination;
First search part, its be based on first key search described in electronic health record database, obtain the first retrieval result;
Whether judging part, the number that it judges first retrieval result is 0, if 0, then returns to prompt message, if
It is not 0 input according to user, judges whether first retrieval result is qualified;
Calculating part, when the judging part judges that first retrieval result is unqualified, sets up the pass of the multiple electronic health record
Key word two-dimensional matrix, and based on the keyword two-dimensional matrix calculate in the multiple electronic health record except first keyword it
The degree of correlation of outer keyword and first keyword, the keyword two-dimensional matrix is with the multiple electronic health record
Keyword is row name, and the medical record number with the multiple electronic health record is to go the binaryzation matrix of name, wherein when in electronic health record
When having the keyword that row name is enumerated, the value of its place ranks in the keyword two-dimensional matrix is designated as 1, be otherwise designated as
0;
In recommended unit, its multiple electronic health record being calculated according to the calculating part in addition to first keyword
Keyword and first keyword the degree of correlation, recommend the maximum keyword of the degree of correlation as the second keyword;
Second search part, its be based on second key search described in electronic health record database, obtain the second retrieval result;
Display part, it can show to the prompt message, first retrieval result and second retrieval result.
2. the retrieval device of electronic medical record system is based on as claimed in claim 1, it is characterised in that
The judging part judges whether the number of first keyword is single;
When the judging part judges first keyword for single, the calculating part is based on following formula and calculates the multiple
The degree of correlation R of keyword in addition to first keyword in electronic health record and first keywordkf,
Wherein, p represents that electronic health record numbers be expert at sequence number, p=1,2,3...., j;J represents the electronics in electronic medical record system
The total number of case history, j is positive integer;K represent the first search key where row sequence number;F is represented except the first keyword
The sequence number of keyword column in addition, f=1 ... k-1, k+1 ... i;I is positive integer, as all keys in electronic medical recordses system
The total number of word (without repeating);Z is the value corresponding to the corresponding ranks in the keyword two-dimensional matrix.
3. the retrieval device of electronic medical record system is based on as claimed in claim 1, it is characterised in that
The judging part judges whether the number of first keyword is single;
When the judging part judges first keyword for two or more word, the calculating part is calculated based on following formula
The degree of correlation Ln of keyword in addition to first keyword in the multiple electronic health record and first keyword,
Wherein, p represents that electronic health record numbers be expert at sequence number, p=1,2,3...., j;J represents the electronics in electronic medical record system
The total number of case history, j is positive integer;A, b, c...t be the first keyword each where row sequence number, n is except first is crucial
The sequence number of the keyword column beyond word.
4. the retrieval device of electronic medical record system is based on as claimed in claim 1, it is characterised in that
When the keyword in the multiple electronic health record that the calculating part is calculated in addition to first keyword with
There is same value in the degree of correlation of first keyword and when being all maximum value, it is same to calculate the degree of correlation respectively based on following formula
The weights We of each keyword column of value,
Wherein, p represents that electronic health record numbers be expert at sequence number, p=1,2,3...., j;J represents the electronics disease in electronic medical record system
The total number gone through, j is positive integer;E represents that the degree of correlation produces the sequence number with the keyword column of value,
It is the second keyword that the recommended unit recommends the maximum keywords of weights We, if the degree of correlation is with each keyword institute of value
Also it is with value, then to select the keyword that any one degree of correlation is maximum value as the second keyword in the weights We of row.
5. a kind of retrieval mode based on electronic health record database, the electronic health record database is included through the multiple electricity of Chinese word segmentation
Sub- case history, it is characterised in that the search method includes:
Input step, the first keyword for being input into user's determination;
First searching step, its be based on first key search described in electronic health record database, obtain the first retrieval result;
Judge step, whether the number that it judges first retrieval result is 0, if 0, then return to prompt message, such as
Fruit is not 0 input according to user, judges whether first retrieval result is qualified;
Calculation procedure, when the judgement step judges that first retrieval result is unqualified, sets up the multiple electronic health record
Keyword two-dimensional matrix, and based on the keyword two-dimensional matrix calculate in the multiple electronic health record except described first is crucial
The degree of correlation of keyword outside word and first keyword, the keyword two-dimensional matrix is with the multiple electronic health record
In keyword be row name, the medical record number with the multiple electronic health record is to go the binaryzation matrix of name, wherein when electronics disease
When having the keyword that row name is enumerated in going through, the value of its place ranks in the keyword two-dimensional matrix is designated as 1, otherwise
It is designated as 0;
In recommendation step, its multiple electronic health record being calculated according to the calculation procedure except first keyword
Outside keyword and first keyword the degree of correlation, recommend the maximum keyword of the degree of correlation as the second keyword;
Second searching step, its be based on second key search described in electronic health record database, obtain the second retrieval result;
Step display, it shows to the prompt message, first retrieval result and second retrieval result.
6. the search method of electronic medical record system is based on as claimed in claim 5, it is characterised in that
It is described to judge that step judges whether the number of first keyword is single;
When it is described judge that step judges first keyword for single when, the calculation procedure is based on following formula and calculates described
The degree of correlation R of keyword in addition to first keyword in multiple electronic health records and first keywordkf,
Wherein, p represents that electronic health record numbers be expert at sequence number, p=1,2,3...., j;J represents the electronics in electronic medical record system
The total number of case history, j is positive integer;K represent the first search key where row sequence number;F is represented except the first keyword
The sequence number of keyword column in addition, f=1 ... k-1, k+1 ... i;I is positive integer, as all keys in electronic medical recordses system
The total number of word (without repeating);Z is the value corresponding to the corresponding ranks in the keyword two-dimensional matrix.
7. the search method of electronic medical record system is based on as claimed in claim 5, it is characterised in that
It is described to judge that step judges whether the number of first keyword is single;
When it is described judge that step judges first keyword for two or more word when, the calculation procedure be based on following formula
Calculate the degree of correlation of the keyword in addition to first keyword and first keyword in the multiple electronic health record
Ln,
Wherein, p represents that electronic health record numbers be expert at sequence number, p=1,2,3...., j;J represents the electronics in electronic medical record system
The total number of case history, j is positive integer;A, b, c...t be the first keyword each where row sequence number, n is except first is crucial
The sequence number of the keyword column beyond word.
8. the search method of electronic medical record system is based on as claimed in claim 5, it is characterised in that
When the keyword in addition to first keyword in the multiple electronic health record that the calculation procedure is calculated
And have same value in the degree of correlation of first keyword and be all maximum value constantly, the degree of correlation is calculated based on following formula respectively
It is the weights We with each keyword column of value,
Wherein, p represents that electronic health record numbers be expert at sequence number, p=1,2,3...., j;J represents the electronics in electronic medical record system
The total number of case history, j is positive integer;E represents that the degree of correlation produces the sequence number with the keyword column of value,
It is the second keyword that the recommendation step recommends the maximum keywords of weights We, if the degree of correlation is with each keyword of value
The weights We of column is also with value, then to select the keyword that any one degree of correlation is maximum value as the second keyword.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510809140.8A CN106776606A (en) | 2015-11-20 | 2015-11-20 | Retrieval device and search method based on electronic health record database |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510809140.8A CN106776606A (en) | 2015-11-20 | 2015-11-20 | Retrieval device and search method based on electronic health record database |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106776606A true CN106776606A (en) | 2017-05-31 |
Family
ID=58885985
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510809140.8A Pending CN106776606A (en) | 2015-11-20 | 2015-11-20 | Retrieval device and search method based on electronic health record database |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106776606A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107818816A (en) * | 2017-12-12 | 2018-03-20 | 刘忠于 | Doctor's electronic health record automatic creation system based on artificial intelligence study |
CN108920896A (en) * | 2018-06-29 | 2018-11-30 | 迈凯基因科技有限公司 | A kind of bone tumour genetic mutation and drug interpret multiple database interactive system |
CN109002449A (en) * | 2017-06-06 | 2018-12-14 | 株式会社日立制作所 | A kind of device and method based on disease association attributes retrieval diagnosed case |
CN109002442A (en) * | 2017-06-06 | 2018-12-14 | 株式会社日立制作所 | A kind of device and method based on doctor's association attributes retrieval diagnosed case |
CN109710670A (en) * | 2018-12-11 | 2019-05-03 | 河南通域医疗科技有限公司 | A method of case history text is converted into structural metadata from natural language |
CN110364234A (en) * | 2019-06-26 | 2019-10-22 | 浙江大学 | Electronic health record intelligent storage analyzing search system and method |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1529865A (en) * | 2000-11-20 | 2004-09-15 | ������ҽ�ƽ�������˾ | System and method for processing patient medical information |
JP2009258870A (en) * | 2008-04-15 | 2009-11-05 | Fuji Xerox Co Ltd | Information extraction device and program, and information extraction rule generation device and program |
US20090327240A1 (en) * | 2007-08-20 | 2009-12-31 | Meehan Stephen W | System And Method For Organizing Data In A Dynamic User-Customizable Interface For Search And Display |
CN102576376A (en) * | 2009-07-08 | 2012-07-11 | 史蒂文·查尔斯·科恩 | Medical history system |
CN103294814A (en) * | 2013-06-07 | 2013-09-11 | 百度在线网络技术(北京)有限公司 | Search result recommendation method, system and search engine |
CN103914797A (en) * | 2013-01-04 | 2014-07-09 | 华为技术有限公司 | Registration scheme recommending system and method |
CN104011765A (en) * | 2012-04-19 | 2014-08-27 | 株式会社东芝 | Medical information search assistance device and medical information search assistance system |
-
2015
- 2015-11-20 CN CN201510809140.8A patent/CN106776606A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1529865A (en) * | 2000-11-20 | 2004-09-15 | ������ҽ�ƽ�������˾ | System and method for processing patient medical information |
US20090327240A1 (en) * | 2007-08-20 | 2009-12-31 | Meehan Stephen W | System And Method For Organizing Data In A Dynamic User-Customizable Interface For Search And Display |
JP2009258870A (en) * | 2008-04-15 | 2009-11-05 | Fuji Xerox Co Ltd | Information extraction device and program, and information extraction rule generation device and program |
CN102576376A (en) * | 2009-07-08 | 2012-07-11 | 史蒂文·查尔斯·科恩 | Medical history system |
CN104011765A (en) * | 2012-04-19 | 2014-08-27 | 株式会社东芝 | Medical information search assistance device and medical information search assistance system |
CN103914797A (en) * | 2013-01-04 | 2014-07-09 | 华为技术有限公司 | Registration scheme recommending system and method |
CN103294814A (en) * | 2013-06-07 | 2013-09-11 | 百度在线网络技术(北京)有限公司 | Search result recommendation method, system and search engine |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109002449A (en) * | 2017-06-06 | 2018-12-14 | 株式会社日立制作所 | A kind of device and method based on disease association attributes retrieval diagnosed case |
CN109002442A (en) * | 2017-06-06 | 2018-12-14 | 株式会社日立制作所 | A kind of device and method based on doctor's association attributes retrieval diagnosed case |
CN107818816A (en) * | 2017-12-12 | 2018-03-20 | 刘忠于 | Doctor's electronic health record automatic creation system based on artificial intelligence study |
CN108920896A (en) * | 2018-06-29 | 2018-11-30 | 迈凯基因科技有限公司 | A kind of bone tumour genetic mutation and drug interpret multiple database interactive system |
CN109710670A (en) * | 2018-12-11 | 2019-05-03 | 河南通域医疗科技有限公司 | A method of case history text is converted into structural metadata from natural language |
CN110364234A (en) * | 2019-06-26 | 2019-10-22 | 浙江大学 | Electronic health record intelligent storage analyzing search system and method |
CN110364234B (en) * | 2019-06-26 | 2022-02-18 | 浙江大学 | Intelligent storage, analysis and retrieval system and method for electronic medical records |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106776606A (en) | Retrieval device and search method based on electronic health record database | |
Das et al. | Group decision making in medical system: An intuitionistic fuzzy soft set approach | |
US20180322954A1 (en) | Method and device for constructing medical knowledge graph and assistant diagnosis method | |
Chen et al. | Artificial intelligence in action: addressing the COVID-19 pandemic with natural language processing | |
CN110299209B (en) | Similar medical record searching method, device and equipment and readable storage medium | |
CN109119134A (en) | Medical history data processing method, medical data recommender system, equipment and medium | |
US20050010444A1 (en) | System and method for assisting medical diagnosis using an anatomic system and cause matrix | |
CN108122611B (en) | Information recommendation method and device, storage medium and program product | |
CN104636430A (en) | Case knowledge base representation and case similarity obtaining method and system | |
Alonso et al. | Evaluation of semantic similarity metrics applied to the automatic retrieval of medical documents: An UMLS approach | |
CN104462554B (en) | Question and answer page relevant issues recommended method and device | |
Ahmad et al. | Intelligent hepatitis diagnosis using adaptive neuro-fuzzy inference system and information gain method | |
Patra et al. | Risk analysis in diabetes prediction based on a new approach of ranking of generalized trapezoidal fuzzy numbers | |
US20200058408A1 (en) | Systems, methods, and apparatus for linking family electronic medical records and prediction of medical conditions and health management | |
Momena et al. | Prediagnosis of disease based on symptoms by generalized dual hesitant hexagonal fuzzy multi-criteria decision-making techniques | |
Chen et al. | Variable selection for joint models of multivariate longitudinal measurements and event time data | |
Rodríguez-González et al. | Analysis of a multilevel diagnosis decision support system and its implications: a case study | |
US20160004829A1 (en) | Clinical laboratory decision support system | |
CN104462556B (en) | Question and answer page relevant issues recommend method and apparatus | |
Chen et al. | Using data mining strategies in clinical decision making: a literature review | |
CN112307028B (en) | Cross-data information knowledge modal differential content recommendation method oriented to essential computation | |
Yaşar et al. | A novel study to increase the classification parameters on automatic three-class COVID-19 classification from CT images, including cases from Turkey | |
Rindskopf | The use of latent class analysis in medical diagnosis | |
CN107256222A (en) | Electronic health record quick retrieval system based on free word retrieval | |
Ling et al. | A matching framework for modeling symptom and medication relationships from clinical notes |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170531 |
|
RJ01 | Rejection of invention patent application after publication |