JP7234622B2 - Document retrieval program, document retrieval method and document retrieval system - Google Patents

Description

The present invention relates to a document search program, document search method and document search system.

2. Description of the Related Art Conventionally, when treating a patient, a doctor sometimes considers a future treatment policy by referring to past medical records. For example, for a patient who has a long history of hospital visits, a doctor may refer to the patient's past medical record to determine the treatment method or drug to be selected. A keyword search, for example, is used to search for medical charts.

As a prior art, for example, there is a technique that performs negative/positive judgment on contact information based on positive words and negative words stored in a storage unit that stores a list of positive words and a list of negative words. In addition, it analyzes whether the classified elements are positive or negative, accumulates the elements in positive and negative dictionaries, compares the elements in the positive dictionary and the elements in the negative dictionary, and determines if they are the same. , there are techniques to remove the same elements from the positive and negative dictionaries.

JP 2016-212925 A JP 2008-204355 A

However, with the conventional technology, there are cases in which it is not possible to determine the positive or negative connotations of words when searching for documents. If it is not possible to determine the positive or negative connotations of words, it is difficult to perform a document search that takes positive or negative conditions into consideration.

In one aspect, the present invention aims to make it possible to identify the tendency of positive or negative connotations of words, and to realize a document search taking positive or negative conditions into consideration.

In one embodiment, with reference to a first storage unit in which documents that can be classified as positive or negative are registered, for a word to be processed, the frequency of appearance of the word in each of the documents that can be classified as positive or negative. A document search program is provided for calculating an index value indicating the tendency of the positive or negative connotation of the word based on the above, and registering the calculated index value in a second storage unit in association with the word. .

According to one aspect of the present invention, it is possible to identify the tendency of positive or negative connotations of words, and to realize document retrieval that takes positive or negative conditions into consideration.

FIG. 1 is an explanatory diagram showing an example of a document retrieval method according to an embodiment. FIG. 2 is an explanatory diagram showing a system configuration example of the document search system 200. As shown in FIG. FIG. 3 is a block diagram showing a hardware configuration example of the information processing apparatus 101. As shown in FIG. FIG. 4 is an explanatory diagram showing an example of the contents stored in the disease name DB 220. As shown in FIG. FIG. 5 is an explanatory diagram showing an example of the contents stored in the chart DB 230. As shown in FIG. FIG. 6 is an explanatory diagram showing an example of the contents stored in the negative/positive DB 240. As shown in FIG. FIG. 7 is a block diagram showing a functional configuration example of the information processing apparatus 101. As shown in FIG. FIG. 8 is an explanatory diagram showing an example of the contents stored in the negative/positive number table 800. As shown in FIG. FIG. 9 is an explanatory diagram showing an example of screen transition of the client device 201. As shown in FIG. FIG. 10 is an explanatory diagram showing another screen example of the search result screen. FIG. 11 is an explanatory diagram showing another screen example of the keyword search screen. FIG. 12 is a flowchart (part 1) showing an example of a negative-positive registration processing procedure of the information processing apparatus 101 . FIG. 13 is a flowchart (No. 2) showing an example of the negative-positive registration processing procedure of the information processing apparatus 101 . FIG. 14 is a flowchart (No. 3) showing an example of the negative-positive registration processing procedure of the information processing apparatus 101 . FIG. 15 is a flow chart showing an example of a specific processing procedure of negative number calculation processing. FIG. 16 is a flow chart showing an example of a specific processing procedure of the number-of-positives calculation process. FIG. 17 is a flowchart (Part 1) showing an example of the document search processing procedure of the information processing apparatus 101 . FIG. 18 is a flowchart (part 2) showing an example of the document search processing procedure of the information processing apparatus 101 .

Exemplary embodiments of a document search program, a document search method, and a document search system according to the present invention will be described below in detail with reference to the drawings.

(Embodiment)
FIG. 1 is an explanatory diagram showing an example of a document retrieval method according to an embodiment. In FIG. 1, an information processing apparatus 101 is a computer that supports document retrieval. Documents are electronic information that can be searched by keywords. Keyword search is a function of searching for documents corresponding to a keyword, and searches for documents containing characters or character strings input as keywords.

Documents to be searched include, for example, articles on medical charts (medical records) created in medical institutions, reports created in research and development, and the like. The patient's chief complaint, doctor's findings (evaluation), diagnosis results, treatment plan (treatment policy), and the like are described in the chart. A research and development report includes the researcher's observations and research results.

Some of these documents have positive content, and some have negative content. Positive means, for example, affirmative, positive, forward-looking, improvement, and the like. Negative means negative, negative, negative, aggravating, and so on. For example, medical record articles include articles (positive) when symptoms have improved and articles (negative) when symptoms have worsened.

Here, in searching for documents, there are cases where it is desired to perform document searches with negative or positive conditions added. For example, when treating a patient with symptoms of "fever", doctors want past medical record articles when the patient's symptoms worsened in order to judge unsuitable treatment methods and drugs for the patient. Sometimes.

However, in document retrieval using keyword search, since the search is performed simply by whether or not the character string entered as the keyword matches, the originally desired information may not be hit, or it may take a long time to find the desired information. There is For example, if a document search is performed using the keyword "fever", the information "fever has subsided", which is not originally necessary, may be hit, and the desired information may be buried.

Therefore, it is conceivable to determine the positive or negative connotation of the document containing the word based on the tendency of the word to have a positive or negative connotation, and perform a document search with the negative or positive condition added. In this case, information specifying tendencies of positive or negative connotations is stored in advance for various words.

However, manually confirming the meanings of various words one by one and registering information indicating tendencies of positive or negative meanings in a database or the like in advance is time-consuming and laborious and unrealistic. Also, depending on the word, it is difficult to manually determine the tendency of positive or negative connotations.

Therefore, in the present embodiment, a document retrieval method that makes it possible to specify the tendency of positive or negative connotations of words and realizes document retrieval that takes into account positive or negative conditions will be described. A processing example of the information processing apparatus 101 will be described below.

(1) The information processing device 101 receives a word to be processed. Here, the word to be processed is a word for which an index value indicating the tendency of the word to have positive or negative connotations is calculated. The positive connotation tendency indicates how positive the connotation is. The tendency of negative connotation represents the degree of negative connotation.

Specifically, for example, the information processing apparatus 101 may accept a word to be processed by a user's operation input or by receiving it from an external computer. Alternatively, the information processing apparatus 101 may extract words from a document and accept the extracted words as words to be processed.

In the example of FIG. 1, it is assumed that the word "xxx" is accepted as the word to be processed.

(2) The information processing apparatus 101 refers to the first storage unit 110, and determines whether the word to be processed is positive or negative based on the frequency of occurrence of the word in each document that can be classified as positive or negative. Calculate an index value that indicates the tendency of meaning.

A first storage unit 110 is a storage unit in which documents that can be classified as positive or negative are registered. A document that can be classified as positive or negative is a document that can be classified as having positive content and results or negative content and results. The first storage unit 110 stores, for example, information for classifying whether the document is positive or negative in association with the document.

Here, it can be said that a word with a higher appearance frequency in positively classified documents has a stronger positive connotation. On the other hand, it can be said that a word with a higher appearance frequency in negatively classified documents has a stronger negative connotation. However, words that appear in positively classified documents may also appear in negatively classified documents. Also, words that appear in negatively classified documents may appear in positively classified documents.

Therefore, the information processing apparatus 101 calculates the index value of a word to be processed based on the appearance frequency in positively classified documents and the appearance frequency in negatively classified documents. The appearance frequency of a word in a document can be obtained, for example, by dividing the number of times the word appears in the document by the total number of words included in the document (total number of words).

In the example of FIG. 1, the first storage unit 110 stores a document 111 classified as positive and a document 112 classified as negative. Also, let the appearance frequency of the word "xxx" in the positively classified documents 111 be "appearance frequency f1" and the appearance frequency of the word "xxx" in the negatively classified documents 112 be "appearance frequency f2".

In this case, the information processing apparatus 101 calculates the index of the word "xxx" based on the appearance frequency f1 in the positively classified document 111 and the appearance frequency f2 in the negatively classified document 112 for the word "xxx". Calculate the value v1. More specifically, for example, the index value v1 may be represented by the difference between the appearance frequency f1 and the appearance frequency f2 (eg, f1-f2).

Here, the index value v1 is a value obtained by subtracting the appearance frequency f2 from the appearance frequency f1. In this case, it can be said that the larger the index value v1, the stronger the tendency of the word "xxx" to have a positive connotation. Also, it can be said that the smaller the index value v1, the stronger the tendency of the word "xxx" to have a negative connotation.

Here, the case where there is one document classified positively (or negatively) has been described as an example, but when there are a plurality of documents, the information processing apparatus 101 determines the number of words in all of the plurality of documents. You may decide to calculate appearance frequency.

(3) The information processing device 101 registers the calculated index value in the second storage unit 120 in association with the word. Here, the second storage unit 120 is a storage unit that stores an index value indicating the tendency of the word to have a positive or negative connotation in association with the word.

In the example of FIG. 1, the information processing apparatus 101 registers the calculated index value v1 in the second storage unit 120 in association with the word "xxx".

As described above, according to the information processing apparatus 101, it is possible to obtain and register an index value that can identify the tendency of the positive or negative connotations of words to be processed. As a result, it becomes possible to estimate the tendency of positive or negative connotations of documents containing the words from the tendency of positive or negative connotations of the words. can be done.

In the example of FIG. 1, an index value v1 that can identify the positive or negative connotation tendency of the word "xxx" can be registered. As a result, when performing a document search, it is possible to use the index value v1 of the word "xxx" to estimate the tendency of positive or negative connotations of documents containing the word "xxx".

(System Configuration Example of Document Search System 200)
Next, an example system configuration of the document retrieval system 200 including the information processing apparatus 101 shown in FIG. 1 will be described. The document retrieval system 200 is applied, for example, to an electronic medical record system of a medical institution. Medical institutions include hospitals, clinics, and the like.

FIG. 2 is an explanatory diagram showing a system configuration example of the document search system 200. As shown in FIG. In FIG. 2 , the document search system 200 includes an information processing device 101 and a client device 201 . In the document search system 200, the information processing device 101 and the client device 201 are connected via a wired or wireless network 210. FIG. The network 210 is, for example, a LAN (Local Area Network), a WAN (Wide Area Network), the Internet, or the like.

Here, the information processing apparatus 101 has a disease name DB (Database) 220 , a chart DB 230 and a negative/positive DB 240 . The information processing device 101 is, for example, a server of an electronic medical record system. The storage contents of various DBs 220, 230, and 240 will be described later with reference to FIGS. 4 to 6. FIG.

A client device 201 is a computer used by a user of the document retrieval system 200 . A user of the document retrieval system 200 is, for example, a medical worker such as a doctor or a nurse at a medical institution. The client device 201 is, for example, a PC (Personal Computer), a tablet PC, or the like.

In the above description, the information processing apparatus 101 and the client apparatus 201 are provided separately, but the present invention is not limited to this. For example, the information processing device 101 may be implemented by the client device 201 .

(Hardware Configuration Example of Information Processing Device 101)
FIG. 3 is a block diagram showing a hardware configuration example of the information processing apparatus 101. As shown in FIG. 3, an information processing apparatus 101 includes a CPU (Central Processing Unit) 301, a memory 302, a disk drive 303, a disk 304, a communication I/F (Interface) 305, and a portable recording medium I/F 306. , and a portable recording medium 307 . Also, each component is connected by a bus 300 .

Here, the CPU 301 controls the entire information processing apparatus 101 . The CPU 301 may have multiple cores. The memory 302 has, for example, a ROM (Read Only Memory), a RAM (Random Access Memory), a flash ROM, and the like. Specifically, for example, a flash ROM stores an OS (Operating System) program, a ROM stores application programs, and a RAM is used as a work area for the CPU 301 . A program stored in the memory 302 is loaded into the CPU 301 to cause the CPU 301 to execute coded processing.

The disk drive 303 controls data read/write with respect to the disk 304 under the control of the CPU 301 . The disk 304 stores data written under the control of the disk drive 303 . Examples of the disk 304 include a magnetic disk and an optical disk.

The communication I/F 305 is connected to the network 210 through a communication line, and is connected to an external computer (for example, the client device 201 shown in FIG. 2) via the network 210 . A communication I/F 305 serves as an interface between the network 210 and the inside of the apparatus, and controls input/output of data from an external computer. For the communication I/F 305, for example, a modem or a LAN adapter can be adopted.

A portable recording medium I/F 306 controls reading/writing of data from/to a portable recording medium 307 under the control of the CPU 301 . The portable recording medium 307 stores data written under control of the portable recording medium I/F 306 . Examples of the portable recording medium 307 include CD (Compact Disc)-ROM, DVD (Digital Versatile Disk), USB (Universal Serial Bus) memory, and the like.

Note that the information processing apparatus 101 may include, for example, an SSD (Solid State Drive), an input device, a display, etc., in addition to the components described above. Further, the information processing apparatus 101 does not have to include, for example, the disk drive 303, the disk 304, the portable recording medium I/F 306, and the portable recording medium 307 among the components described above. Also, the client device 201 shown in FIG. 2 can be realized by a hardware configuration similar to that of the information processing device 101 . However, the client device 201 has an input device, a display, etc. in addition to the components described above.

(Stored contents of various DBs 220, 230, 240)
Next, the storage contents of the various DBs 220, 230, and 240 of the information processing apparatus 101 will be described with reference to FIGS. 4 to 6. FIG. Various DBs 220, 230, and 240 are realized by storage devices such as the memory 302 and the disk 304 shown in FIG. 3, for example.

FIG. 4 is an explanatory diagram showing an example of the contents stored in the disease name DB 220. As shown in FIG. 4, the disease name DB 220 has fields of patient ID, disease name code, disease name, main disease name category, disease name start date, outcome, and outcome date. Information 400-1 to 400-8) are stored as records.

Here, the patient ID is an identifier that uniquely identifies a patient. A patient is, for example, a patient at a medical institution where the document retrieval system 200 is introduced. A disease name code is an identifier that uniquely identifies a disease name given to a patient. A disease name is a name of a disease. The name of the disease includes the name of the primary disease, the name of the suspected disease, and the like. The main disease name is the main disease name. A suspected disease name is a suspected disease name.

The main disease name classification is a flag indicating whether or not it is the main disease name. The main disease name category "1" indicates that it is the main disease name. The main disease name category "0" indicates that it is not the main disease name. The disease name start date is the date when the disease name was given. Outcome indicates what kind of condition the patient has developed as a result of progression or improvement of the disease. Outcome date is the date the outcome was described.

For example, the disease name information 400-1 includes the disease name code "12345", the disease name "stomach tumor", the main disease category "1", the disease name start date "2018/10/1", and the outcome "death" of the patient with the patient ID "P1". ” and the outcome date “2018/11/1”.

FIG. 5 is an explanatory diagram showing an example of the contents stored in the chart DB 230. As shown in FIG. 5, the medical record DB 230 has patient ID, entry date, SOAP category and article fields, and by setting information in each field, medical record information (eg, medical record information 500-1 to 500-8) can be obtained. Store as a record.

Here, the patient ID is an identifier that uniquely identifies a patient. The entry date is the date when the article in the medical chart was entered. The SOAP section indicates which information in the medical record described in SOAP format is the article. The SOAP category "S" indicates the chief complaint of the patient. "S" is an acronym for Subjective.

The SOAP section “O” indicates patient test results and the like. "O" is an acronym for Objective. The SOAP section "A" indicates the doctor's findings (evaluation). "A" is an acronym for Assessment. SOAP section "P" indicates a treatment plan. "P" is an acronym for Plan. The article is the description content of the chart.

For example, the chart information 500-1 indicates patient ID "P1", entry date "2018/10/1", SOAP category "S", and article "lower abdominal pain since 1 week ago".

FIG. 6 is an explanatory diagram showing an example of the contents stored in the negative/positive DB 240. As shown in FIG. 6, the negative-positive DB 240 has fields of element, disease name code, disease name, positive number, positive appearance frequency, negative number, negative appearance frequency, and negative-positive frequency. By setting information in each field, negative-positive information (for example, negative-positive information 600-1 to 600-10) is stored as a record.

Here, the elements are the words used in the document. The document is, for example, an article of medical record information. A disease name code is an identifier that uniquely identifies a disease name. A disease name is a name of a disease. The positive count is the number of times the word appears in positively classified documents. The positive frequency of occurrence indicates the frequency of occurrence of the word in positively classified documents.

Negative count is the number of times a word appears in documents classified as negative. Negative frequency indicates the frequency of occurrence of words in negatively classified documents. The negative-positive frequency is an index value that indicates the tendency of a word to have a positive or negative connotation. Here, the negative-positive frequency indicates a value obtained by subtracting the negative appearance frequency from the positive appearance frequency.

For example, the negative/positive information 600-1 includes the element "pain", the disease name code "12345", the disease name "stomach tumor", the positive number "5", the positive appearance frequency "0.113636364", the negative number "10", and the negative appearance frequency. "0.263157895" and negative-positive power "-0.149521531" are shown.

(Example of functional configuration of information processing apparatus 101)
FIG. 7 is a block diagram showing a functional configuration example of the information processing apparatus 101. As shown in FIG. 7, the information processing apparatus 101 includes a first reception unit 701, a calculation unit 702, a registration unit 703, a second reception unit 704, a search unit 705, a determination unit 706, and an output unit 707. , a first storage unit 710 and a second storage unit 720 . Specifically, for example, the first reception unit 701 to the output unit 707 cause the CPU 301 to execute a program stored in a storage device such as the memory 302 and the disk 304 shown in FIG. /F305 realizes the function. The processing results of each functional unit are stored in a storage device such as the memory 302 or disk 304, for example. The first storage unit 710 and the second storage unit 720 are realized by storage devices such as the memory 302 and the disk 304, for example.

The first storage unit 710 stores documents that can be classified as positive or negative. Specifically, for example, the first storage unit 710 stores the disease name DB 220 shown in FIG. 4 and the chart DB 230 shown in FIG. The first storage unit 110 illustrated in FIG. 1 corresponds to the first storage unit 710, for example.

The second storage unit 720 stores an index value indicating the tendency of the word to have a positive or negative connotation in association with the word. Specifically, for example, the second storage unit 720 stores the negative-positive DB 240 shown in FIG. The second storage unit 120 illustrated in FIG. 1 corresponds to the second storage unit 720, for example.

A first reception unit 701 receives a word to be processed. Here, the word to be processed is a word for which an index value indicating the tendency of the word to have positive or negative connotations is calculated. The words to be processed may be words of a particular part of speech. A specific part of speech can be arbitrarily set, and includes, for example, nouns, verbs, adjectives, adverbs, and adverbs.

Specifically, for example, the first reception unit 701 receives words extracted from a document as words to be processed. The document may be, for example, a positive or negative classifiable document stored in the first storage unit 710 . More specifically, for example, the first reception unit 701 receives words of a specific part of speech extracted from articles of medical record information in the medical record DB 230 shown in FIG. 5 as words to be processed. good too.

The calculation unit 702 refers to the first storage unit 710, and for the word to be processed, based on the appearance frequency of the word in each document that can be classified as positive or negative, determines the tendency of the word to have a positive or negative connotation. Calculates an index value indicating

In the following description, an index value indicating a tendency of positive or negative connotations may be referred to as "negative-positive frequency".

Specifically, for example, the calculation unit 702 classifies each document stored in the first storage unit 710 as either positive or negative. Next, the calculation unit 702 calculates the appearance frequency of words in positively classified documents. The calculation unit 702 also calculates the frequency of appearance of words in negatively classified documents. Then, the calculating unit 702 calculates the negative-positive frequency of the word based on the calculated appearance frequency of the word in the positively classified document and the calculated appearance frequency of the word in the negatively classified document.

More specifically, for example, the calculation unit 702 selects disease name information from the disease name DB 220 shown in FIG. Next, the calculation unit 702 specifies a period T from the disease name start date to the outcome date of the selected disease name information. Then, the calculation unit 702 selects medical chart information that corresponds to the patient ID of the selected disease name information and that includes the date of description in the specified period T from the medical chart DB 230 shown in FIG.

Here, if the outcome of the selected disease name information is "death", the calculation unit 702 classifies the article of the selected medical chart information as negative. On the other hand, when the outcome of the selected disease name information is "remission or cure", the calculation unit 702 classifies the article of the selected medical chart information as positive.

This allows medical chart information articles written during treatment of an illness to be classified as positive or negative, depending on whether the patient's outcome is "remission or cure" or "death". can.

Of the disease name information in the disease name DB 220, only disease name information whose disease name start date is within a predetermined period may be processed. The predetermined period can be set arbitrarily, and is set to a period such as the last five years, for example. This makes it possible to exclude articles of old medical record information from being processed.

Further, among the disease name information in the disease name DB 220, only the disease name information whose disease name is the main disease name, that is, only the disease name information of the main disease name classification "1" may be processed. As a result, for example, medical record information articles written when the name of the disease has not been determined can be excluded from the processing targets.

Next, the calculation unit 702 morphologically analyzes the article of the medical record information classified as positive, thereby breaking down the article into words (morphemes). Next, the calculation unit 702 extracts words of a specific part of speech from the decomposed words (morphemes). Then, the calculation unit 702 calculates the number of appearances of each extracted word in articles of medical record information classified as positive.

The number of appearances in articles of medical record information classified as positive may be calculated, for example, for all articles of medical record information among articles of medical record information classified positively. It may be calculated for articles of medical record information.

In addition, the calculation unit 702 performs morphological analysis on articles of medical record information classified as negative, thereby decomposing the articles into word (morpheme) units. Next, the calculation unit 702 extracts words of a specific part of speech from the decomposed words (morphemes). Then, the calculation unit 702 calculates the number of appearances of each extracted word in articles of medical record information classified as negative.

The number of appearances in articles of medical record information classified as negative may be calculated, for example, for all articles of medical record information among the articles of medical record information classified as negative. It may be calculated for articles of medical record information.

In the following description, the number of occurrences of a word in positively classified documents (for example, articles of medical chart information) may be referred to as "positive number". Also, the number of occurrences of a word in a negatively classified document (for example, an article in medical record information) may be expressed as "negative number".

The positive counts and negative counts of the calculated words are stored, for example, in a negative-positive count table 800 as shown in FIG. The negative/positive number table 800 is realized by a storage device such as the memory 302 and the disk 304, for example.

FIG. 8 is an explanatory diagram showing an example of the contents stored in the negative/positive number table 800. As shown in FIG. In FIG. 8, the negative/positive number table 800 has fields for element, disease name code, disease name, negative number, and positive number. -6) is stored as a record.

Here, the elements are the words used in the medical record information article. A disease name code is an identifier that uniquely identifies a disease name. The disease name is specified from the disease name information corresponding to the medical chart information, ie, the disease name information referred to when selecting the medical chart information. A disease name is a name of a disease. The number of negatives is the number of occurrences of a word in articles of medical record information classified as negative. The number of positives is the number of occurrences of a word in articles of medical record information classified as positive.

For example, the number information 800-1 indicates the element "pain", disease name code "12345", disease name "stomach tumor", negative number "5" and positive number "10".

For example, the calculation unit 702 refers to the negative/positive number table 800 and calculates the total number of positives by accumulating the number of positives for each word. The total number of positive cases corresponds to the total number of words (for example, words of a specific part of speech) included in articles of medical record information classified as positive. Next, the calculation unit 702 calculates the frequency of occurrence of positives for each word by dividing the number of positives for each word by the total number of positives. The positive appearance frequency is the appearance frequency of a word in articles of medical record information classified as positive.

The calculation unit 702 also refers to the negative/positive number table 800 and calculates the total number of negatives by accumulating the number of negatives of each word. The total number of negative cases corresponds to the total number of words (for example, words of a specific part of speech) included in articles of medical record information classified as negative. Next, the calculation unit 702 calculates the negative appearance frequency of each word by dividing the number of negative occurrences of each word by the total number of negative occurrences. Negative appearance frequency is the appearance frequency of a word in articles of medical record information classified as negative.

Then, the calculation unit 702 calculates the negative-positive frequency of each word based on the calculated positive appearance frequency and negative appearance frequency of each word. For example, the calculation unit 702 may calculate the difference between the positive appearance frequency and the negative appearance frequency of each word as the negative-positive frequency of each word.

Assuming that the negative-positive frequency is a value obtained by subtracting the negative-appearance frequency from the positive frequency, it can be said that the higher the negative-positive frequency, the stronger the tendency of the word to have a positive connotation. On the other hand, if the negative-positive frequency is a value obtained by subtracting the positive-appearance frequency from the negative frequency, it can be said that the larger the negative-positive frequency, the stronger the negative connotation of the word. Here, the negative-positive frequency is a value obtained by subtracting the negative appearance frequency from the positive appearance frequency.

The registration unit 703 registers the calculated negative-positive frequency of the word in the second storage unit 720 in association with the word. Specifically, for example, the registration unit 703 registers the calculated negative-positive frequency of the word in the negative-positive DB 240 shown in FIG. 6 in association with the word (element). At this time, the registration unit 703 further associates the disease name code, disease name, positive number, positive appearance frequency, negative number, and negative appearance frequency, and registers them in the negative/positive DB 240 .

The disease name is specified from the disease name information corresponding to the medical record information referred to when calculating the negative-positive frequency of the word. However, the positive number, positive appearance frequency, negative number, and negative appearance frequency may not be registered in the negative/positive DB 240 .

A second reception unit 704 receives a search string. Here, the search character string designates a keyword (character or character string) as a search condition when searching for documents. A search string may include multiple keywords. Specifically, for example, the second accepting unit 704 accepts the received search character string by receiving the search character string from the client device 201 shown in FIG.

More specifically, for example, when a doctor treats a patient, he enters a search character string to search for past medical record articles on a keyword search screen 910 shown in FIG. 9, which will be described later. In this case, the client device 201 transmits the input search character string to the information processing device 101 . The search character string may be associated with information specifying the disease name of the patient to be treated.

A search unit 705 searches for a document corresponding to the received search character string from the document group. Here, the document group is, for example, a document group stored in the first storage unit 710 . Also, the document group may be a document group that can be obtained from a website on the Internet.

Specifically, for example, the search unit 705 searches the medical record DB 230 for an article of medical record information corresponding to the search character string. More specifically, for example, the search unit 705 searches for articles in medical record information that include characters or character strings included in the search character string. At this time, the search unit 705 may search articles of medical record information with the SOAP classification of "S" or "A".

Further, the search unit 705 may search for articles in medical record information describing the treatment of a disease with the same name as the current name of the disease (name of the patient to be diagnosed). A disease name corresponding to medical record information can be specified from the disease name DB 220 based on, for example, the patient ID and the written date of the medical record information. However, any existing search algorithm may be used.

In the following description, articles of medical chart information in the medical chart DB 230 may simply be referred to as "medical chart articles".

When one or more documents corresponding to the search character string are retrieved, the determination unit 706 determines the priority of each of the retrieved one or more documents. Here, the priority of a document is the priority when outputting a document as a search result for a search character string. For example, a document with a higher priority is displayed higher than other documents in the search results.

Specifically, for example, the determination unit 706 refers to the second storage unit 720 and determines the negative-positive value of each of the retrieved one or more documents based on the negative-positive frequency corresponding to the word included in each document. Calculate the frequency. The determining unit 706 also refers to the second storage unit 720 and calculates the negative-positive frequency of the search character string based on the negative-positive frequency corresponding to the word included in the search character string. Then, the determination unit 706 determines the priority of each document based on the calculated negative-positive frequency of each document and the negative-positive frequency of the search character string.

More specifically, for example, the determining unit 706 morphologically analyzes the retrieved medical record article to break it down into words (morphemes). Next, the determining unit 706 extracts words of a specific part of speech from the decomposed words (morphemes).

The determination unit 706 also identifies the disease name of the patient to be treated. The disease name of the patient to be treated may be identified from information associated with the search character string, or may be identified by querying the client device 201 or the electronic medical record system.

Next, the determination unit 706 refers to the negative-positive DB 240 for each extracted word, and specifies the negative-positive frequency corresponding to the specified disease name. That is, the determining unit 706 identifies the negative-positive frequency corresponding to the combination of the word and the disease name. For example, the negative-positive frequency corresponding to the combination of the word "pain" and the disease name "stomach tumor" is "-0.149521531".

Then, the determining unit 706 calculates the negative-positive frequency of the medical record article by accumulating the negative-positive frequency of each specified word. At this time, the determining unit 706 may normalize the sum of the negative-positive frequencies of each word by dividing it by the number of words extracted from the medical record article (or the number of decomposed words).

As a result, it is possible to obtain an index value (negative-positive frequency) that indicates the tendency of the positive or negative connotation of the medical record article corresponding to the search character string.

The determination unit 706 also performs morphological analysis on the search string to decompose the search string into words (morphemes). Next, the determining unit 706 extracts words of a specific part of speech from the decomposed words (morphemes). Then, the determination unit 706 refers to the negative-positive DB 240 for each extracted word, and specifies the negative-positive frequency corresponding to the disease name of the specified patient to be treated.

Next, the determination unit 706 calculates the negative-positive frequency of the search character string by accumulating the negative-positive frequency of each specified word. At this time, the determination unit 706 may normalize the sum of the negative-positive frequencies of each word by dividing the number of words extracted from the search character string (or the number of decomposed words).

As a result, it is possible to obtain an index value (negative-positive frequency) that indicates the tendency of the search character string to have a positive or negative connotation.

Then, the determination unit 706 determines the priority of each medical chart article based on the calculated negative-positive frequency of each medical chart article and the negative-positive frequency of the search character string. Here, first and second determination examples for determining the priority of medical chart articles based on the negative-positive frequency of each medical chart article and the negative-positive frequency of the search character string will be described.

First Determination Example Based on the negative-positive frequency of each medical chart article and the negative-positive frequency of the search character string, the determination unit 706 selects one or more retrieved medical chart articles whose negative-positive frequency is close to the search character string. The priority of each medical chart article is determined so that the medical chart article has a higher priority. As a result, it is possible to give a higher priority to medical record articles that tend to have positive or negative connotations as strong as the search character string.

At this time, the determining unit 706 may classify the search string as positive or negative based on the calculated negative-positive frequency of the search string. For example, when the negative-positive frequency of the search string is 0 or more, the determining unit 706 classifies the search string as positive. On the other hand, when the negative-positive frequency of the search string is less than 0, the determining unit 706 classifies the search string as negative. However, it is assumed that the larger the value of the negative-positive frequency, the stronger the tendency of the positive connotation.

Here, when the search character string is classified as positive, it can be said that the searcher needs search results with a positive connotation rather than search results with a negative connotation. On the other hand, if the search character string is classified as negative, it can be said that the searcher needs search results with a negative connotation rather than search results with a positive connotation.

Therefore, when the search character string is classified as positive, the determination unit 706 assigns each medical chart article a priority such that a medical chart article with a negative-positive frequency of 0 or more has a higher priority than a medical chart article with a negative-positive frequency of less than 0. may be determined. As a result, when the search character string is classified as positive, the chart article whose negative-positive frequency is 0 or more and which is closer to the search character string can be given higher priority.

On the other hand, if the search character string is classified as negative, the determining unit 706 assigns a higher priority to medical chart articles with a negative-positive frequency of less than 0 than to medical chart articles with a negative-positive frequency of 0 or more. Priority may be determined. As a result, when the search character string is classified as negative, the chart article whose negative-positive count is less than 0 and which is closer to the search character string can be given higher priority.

Second Determination Example The determination unit 706 classifies the search string as positive or negative based on the calculated negative-positive frequency of the search string. Here, when the search string is classified as positive, the determination unit 706 may determine the priority of each medical chart article so that the medical chart article with a higher negative-positive frequency has a higher priority. As a result, when the search character string is classified as positive, the chart article that tends to have a stronger positive connotation can be given higher priority.

On the other hand, if the search character string is classified as negative here, the determination unit 706 may determine the priority of each medical chart article so that the medical chart article with a smaller negative-positive frequency number has a higher priority. . As a result, when the search character string is classified as negative, the chart article that tends to have a stronger negative connotation can be given a higher priority.

It is possible to arbitrarily set how to determine the priority of each medical chart article. For example, the information processing apparatus 101 may determine the priority of each medical record article according to an instruction from the client apparatus 201 by either the first or second determination method described above.

Further, the determining unit 706 may determine the display mode of each medical chart article based on the calculated negative-positive frequency of each medical chart article. Here, the display mode of the medical chart article is the mode when the medical chart article is displayed as a search result, and is represented by the background color, character color, size, and the like of the medical chart article.

Specifically, for example, the determining unit 706 determines the display mode of each medical chart article such that the display mode of a medical chart article with a negative-positive frequency of 0 or more is different from the display mode of a medical chart article with a negative-positive frequency of less than 0. . This makes it possible to distinguish between search results with a positive connotation and search results with a negative connotation.

A display example of medical record articles in the display mode determined by the determination unit 706 will be described later with reference to FIG. 9 .

The output unit 707 outputs search results corresponding to the search character string based on the determined priority of each document. Specifically, for example, the output unit 707 outputs a document list (or a list of document names, etc.) in which the retrieved one or more documents are arranged in descending order of priority as the search result. You can decide.

At this time, the output unit 707 may output, as search results, the top N documents among the retrieved one or more documents in descending order of the determined priority. N can be set arbitrarily. The output unit 707 may output each document (or the name of the document, etc.) as a search result in the determined display mode.

Further, when outputting each document (or the name of the document, etc.), the output unit 707 may output the calculated index value of each document in association with each document. The output format of the output unit 707 includes, for example, transmission to another computer via the communication I/F 305, storage in a storage device such as the memory 302 and disk 304, display on a display (not shown), and output to a printer (not shown). print output, etc.

More specifically, for example, the output unit 707 may display, on the client device 201, a search result screen showing a list of medical record articles corresponding to the search character string. A screen example of the search result screen will be described later with reference to FIGS. 9 and 10. FIG.

In the above description, the search character string is classified into positive or negative based on the negative-positive frequency of the search character string, but the classification is not limited to this. For example, the second reception unit 704 may receive selection of either positive or negative along with the search character string.

Selection of positive or negative is performed, for example, on a keyword search screen 1100 displayed on the client device 201 and shown in FIG. 11 described later. In this case, the client device 201 transmits the input search character string and the positive or negative selection result to the information processing device 101 .

Then, the second reception unit 704 receives the search character string and the selection result from the client device 201, thereby accepting the search character string and the selection of positive or negative. The determining unit 706 classifies the search character string as positive when positive selection is received. On the other hand, when negative selection is accepted, the determining unit 706 classifies the search character string as negative.

Further, in the above description, the negative-positive frequency of each document is calculated for each document when searching for documents, but the present invention is not limited to this. For example, the information processing apparatus 101 may calculate in advance the negative-positive frequency of the document stored in the first storage unit 710 .

Specifically, for example, the information processing apparatus 101 refers to the second storage unit 720, and based on the negative-positive frequency corresponding to the word included in each document stored in the first storage unit 710, each document Calculate the negative-positive frequency of Then, the information processing apparatus 101 may register the calculated negative-positive frequency of each document in the first storage unit 710 in association with each document.

More specifically, for example, the information processing apparatus 101 refers to the negative-positive DB 240 for words included in each medical chart article stored in the medical chart DB 230, and specifies the negative-positive frequency corresponding to the disease name corresponding to each medical chart article. do. Next, the information processing apparatus 101 calculates the negative-positive frequency of each medical chart article based on the negative-positive frequency of words included in each specified medical chart article. Then, the information processing apparatus 101 may register the calculated negative-positive frequency of each medical chart article in a storage device such as the memory 302 and the disk 304 in association with each medical chart article.

As a result, when retrieving documents, it is only necessary to read out the negative-positive frequencies of documents registered in advance (e.g., medical chart articles). response performance can be improved.

Further, the information processing apparatus 101 may obtain the negative-positive frequency of medical chart articles, for example, for medical chart articles written by trainees or the like. As a result, it is possible to evaluate whether the content of the medical record article is too positive or too negative based on the negative-positive frequency of the medical record article, which can be used for training new employees.

Also, each functional unit of the information processing apparatus 101 may be realized by another computer different from the information processing apparatus 101 in the document search system 200, for example, the client apparatus 201. FIG. Also, each functional unit of the information processing apparatus 101 may be implemented by a plurality of computers (for example, the information processing apparatus 101 and the client apparatus 201) in the document search system 200. FIG.

(Screen examples of various screens displayed on the client device 201)
Next, screen examples of various screens displayed on the display (not shown) of the client device 201 will be described. Here, it is assumed that a doctor considers a future treatment policy by referring to the patient's past medical chart articles when treating a patient.

In the following description, as an operation for the user to select a button or the like displayed on the operation screen, a case of performing a click operation will be described as an example.

FIG. 9 is an explanatory diagram showing an example of screen transition of the client device 201. As shown in FIG. In FIG. 9, a keyword search screen 910 is an operation screen for accepting input of a search character string. On the keyword search screen 910, a user (doctor) can input a search character string by clicking a box 911 using an input device (not shown). In the example of FIG. 9, the search string “pain in lower abdomen” is entered in box 911 .

When a user (doctor) clicks a button 912 on the keyword search screen 910 , the search character string input in the box 911 is transmitted to the information processing apparatus 101 . In the example of FIG. 9 , the search character string “pain in lower abdomen” input in box 911 is transmitted to information processing apparatus 101 .

When the information processing apparatus 101 receives the search character string “lower abdominal pain” from the client device 201, it searches the medical record DB 230 for a medical record article corresponding to the search character string “lower abdominal pain”. Next, the information processing apparatus 101 determines the priority of each retrieved medical record article. Here, it is assumed that the priority of each medical record article is determined according to the first determination example described above.

Specifically, for example, the information processing apparatus 101 calculates the negative-positive frequency of the search character string “pain in the lower abdomen”. In addition, the information processing apparatus 101 calculates the negative-positive frequency of each searched medical record article. Here, a case of calculating the negative-positive frequency of the search character string "pain in the lower abdomen" will be taken as an example to describe specific processing contents.

More specifically, for example, the information processing apparatus 101 morphologically analyzes the search character string “pain in the lower abdomen” and breaks it down into words (morphemes). Here, it is decomposed into three words, ``lower abdomen'', ``no'' and ``pain''. Next, the information processing apparatus 101 refers to the negative-positive DB 240 to specify the negative-positive frequency corresponding to each decomposed word.

However, the disease name of the patient is assumed to be "stomach tumor". Also, the word "no" is excluded because it is a particle. That is, the information processing apparatus 101 refers to the negative-positive DB 240 and specifies the negative-positive frequency corresponding to the disease name "stomach tumor" for the word "lower abdomen" and the word "pain" among the three decomposed words.

Here, it is assumed that "-0.017942584" corresponding to the disease name "stomach tumor" is specified as the negative-positive frequency of the word "lower abdomen". It is also assumed that "-0.149521531" corresponding to the disease name "stomach tumor" is specified as the negative-positive frequency of the word "pain".

Next, the information processing apparatus 101 calculates the negative-positive frequency of the search character string “pain in the lower abdomen” by accumulating the negative-positive frequency of each specified word. Here, the negative-positive frequency of the search character string "pain in the lower abdomen" is "-0.167464115". Therefore, the search character string “pain in the lower abdomen” is classified as negative.

In this case, the information processing apparatus 101 assigns a higher priority to a medical chart article having a negative-positive frequency of less than 0 and a negative-positive frequency closer to the search character string among the retrieved medical chart articles. Then, the information processing apparatus 101 displays on the client device 201 a search result screen 920 in which the searched medical chart articles are arranged according to the determined priority of each medical chart article.

The search result screen 920 is an operation screen showing a list of medical record articles 921 to 924 corresponding to the search character string "pain in the lower abdomen". On the search result screen 920, chart articles having a negative-positive frequency of less than 0 and having a negative-positive frequency close to the search character string "lower abdominal pain" are preferentially displayed (displayed at the top).

Here, the negative-positive frequency of the medical record article 921 is "-0.16755522". Also, the negative-positive frequency of the medical record article 922 is "-0.16899999". The negative-positive frequency of the medical record article 923 is "-0.147856321". The negative-positive frequency of the medical record article 924 is "+0.157856561".

For example, the medical record article 921 has a negative-positive frequency of less than 0 and is closest to the search character string "lower abdominal pain", so it is displayed at the top. Further, since the medical chart article 924 has a negative-positive frequency of 0 or more, it is displayed below the other medical chart articles 921 to 923 whose negative-positive frequency is less than zero.

According to the search result screen 920, it is possible to present search results with negative conditions added according to the search character string "pain in the lower abdomen". As a result, the doctor can easily find the necessary information from the patient's past medical records when treating the patient with the disease name "stomach tumor".

In addition, on the search result screen 920, the medical chart articles 921 to 923 with a negative-positive frequency of less than 0 and the medical chart article 924 with a negative-positive frequency of 0 or more are displayed in different manners (different background patterns). This allows the doctor to easily discriminate between medical chart articles with positive connotations and medical chart articles with negative connotations.

Further, on the search result screen 920, negative-positive frequency information 931-934 are displayed in association with medical record articles 921-924, respectively. Each of the negative-positive frequency information 931-934 is information indicating the negative-positive frequency of each medical record article 921-924. This allows the doctor to determine the strength of the positive or negative tendency of each of the medical chart articles 921-924 from the negative-positive frequency of each of the medical chart articles 921-924.

However, the negative-positive frequency information 931 to 934 may not be displayed on the search result screen 920 .

Next, another screen example of the search result screen will be described with reference to FIG. Here, it is assumed that the priority of each medical record article is determined according to the second determination example described above. In this case, if the search character string “pain in the lower abdomen” is classified as negative, the chart article with a stronger negative connotation tends to have a higher priority.

FIG. 10 is an explanatory diagram showing another screen example of the search result screen. In FIG. 10, a search result screen 1000 is an operation screen showing a list of medical record articles 921 to 924 corresponding to the search character string "lower abdominal pain". On the search result screen 1000, medical chart articles that tend to have a strong negative connotation are preferentially displayed (displayed at the top).

As a result, the doctor can see that the higher the displayed information, the stronger the negative connotation, and the doctor can easily find the necessary information from the past medical record articles.

Next, another screen example of the keyword search screen will be described with reference to FIG.

FIG. 11 is an explanatory diagram showing another screen example of the keyword search screen. In FIG. 11, a keyword search screen 1100 is an operation screen for accepting input of a search character string. In the keyword search screen 1100, a user (doctor) can input a search character string by clicking a box 1101 through an operation input. In the example of FIG. 11, a search string “pain in lower abdomen” is entered in box 1101 .

Also, on the keyword search screen 1100, the user can click either button 1102 or 1103 to select either positive or negative. In the example of FIG. 11, button 1103 is clicked to select negative.

When a user clicks a button 1104 on the keyword search screen 1100 , the search character string entered in the box 1101 and the positive or negative selection result are sent to the information processing apparatus 101 . In the example of FIG. 11 , the search character string “lower abdominal pain” input in the box 1101 and the selection result “negative” are transmitted to the information processing apparatus 101 .

As a result, when classifying the search character string "pain in the lower abdomen" into positive or negative in the information processing apparatus 101, the selection result "negative" can be obtained without calculating the negative-positive frequency of the search character string "pain in the lower abdomen". can be classified with high accuracy from

(Negative-Positive Registration Processing Procedure of Information Processing Device 101)
Next, negative/positive registration processing procedures of the information processing apparatus 101 will be described with reference to FIGS. 12 to 14. FIG.

12 to 14 are flow charts showing an example of a negative/positive registration processing procedure of the information processing apparatus 101. FIG. 12, the information processing apparatus 101 first selects unselected disease name information from the disease name DB 220 (step S1201). Next, the information processing apparatus 101 refers to the main disease name category of the selected disease name information and determines whether or not it is the main disease name (step S1202).

Here, if it is not the main disease name (step S1202: No), the information processing apparatus 101 proceeds to step S1208. On the other hand, if it is the main disease name (step S1202: Yes), the information processing apparatus 101 determines whether the disease name start date of the selected disease name information is within five years (step S1203).

Here, if the disease name start date is not within five years (step S1203: No), the information processing apparatus 101 proceeds to step S1208. On the other hand, if the disease name start date is within five years (step S1203: Yes), the information processing apparatus 101 determines whether the outcome of the selected disease name information is "death" (step S1204).

Here, if the outcome is "death" (step S1204: Yes), the information processing apparatus 101 executes negative number calculation processing (step S1205), and proceeds to step S1208. A specific processing procedure of the negative number calculation processing will be described later with reference to FIG. 15 .

On the other hand, if the outcome is not "death" (step S1204: No), the information processing apparatus 101 determines whether the outcome of the selected disease name information is "remission or cure" (step S1206). If the outcome is not "remission or cure" (step S1206: No), the information processing apparatus 101 proceeds to step S1208.

On the other hand, if the outcome is "remission or cure" (step S1206: Yes), the information processing apparatus 101 executes positive number calculation processing (step S1207). A specific processing procedure of the number-of-positives calculation processing will be described later with reference to FIG. 16 . Then, the information processing apparatus 101 determines whether there is unselected disease name information from the disease name DB 220 (step S1208).

If there is unselected disease name information (step S1208: Yes), the information processing apparatus 101 returns to step S1201. On the other hand, if there is no unselected disease name information (step S1208: No), the information processing apparatus 101 proceeds to step S1301 shown in FIG.

In the flowchart of FIG. 13, the information processing apparatus 101 first refers to the negative/positive number table 800 and calculates the total number of negatives by accumulating the number of negatives of each word (step S1301). Next, the information processing apparatus 101 refers to the negative/positive number table 800 and calculates the total number of positives by accumulating the number of positives of each word (step S1302).

Then, the information processing apparatus 101 selects unselected number information that has not been selected from the negative/positive number table 800 (step S1303). Next, the information processing apparatus 101 determines whether or not the number of negatives in the selected number of cases information has not been set (step S1304).

Here, if the number of negatives has already been set (step S1304: No), the information processing apparatus 101 divides the number of negatives in the selected number of cases information by the calculated total number of negatives to calculate the appearance frequency of negatives (step S1305), and proceeds to step S1307.

On the other hand, if the number of negatives is not set (step S1304: Yes), the information processing apparatus 101 sets the frequency of appearance of negatives to "0" (step S1306). Then, the information processing apparatus 101 associates the element of the selected number of cases information with the disease name, stores the negative appearance frequency in the negative/positive DB 240 (step S1307), and proceeds to step S1401 shown in FIG.

In the flowchart of FIG. 14, first, the information processing apparatus 101 determines whether or not the positive number of the selected number of cases information has not been set (step S1401). Here, if the number of positives has already been set (step S1401: No), the information processing apparatus 101 divides the number of positives in the selected number of information by the calculated total number of positives to calculate the appearance frequency of positives (step S1402), and proceeds to step S1404.

On the other hand, if the number of positives has not been set (step S1401: YES), the information processing apparatus 101 sets the appearance frequency of positives to "0" (step S1403). Then, the information processing apparatus 101 stores the positive frequency of appearance in the negative-positive DB 240 in association with the element of the selected number of cases information and the disease name (step S1404).

Next, the information processing apparatus 101 calculates the negative-positive frequency by subtracting the negative appearance frequency from the positive appearance frequency (step S1405). Then, the information processing apparatus 101 stores the calculated negative-positive frequency in the negative-positive DB 240 in association with the element of the selected number of cases information and the disease name (step S1406).

Next, the information processing apparatus 101 determines whether or not there is unselected number information that has not been selected from the negative/positive number table 800 (step S1407). Here, if there is unselected number information (step S1407: Yes), the information processing apparatus 101 returns to step S1303.

On the other hand, if there is no unselected number information (step S1407: No), the information processing apparatus 101 terminates the series of processes according to this flowchart. The information processing apparatus 101 may store the negative number of cases and the positive number of cases of the selected number of cases information in the negative-positive DB 240 in association with the selected number of cases information element and disease name.

Thereby, it is possible to register an index value (negative-positive frequency) indicating the tendency of the words used in the medical record to have positive or negative connotations.

Next, a specific processing procedure of the negative count calculation processing in step S1205 shown in FIG. 12 will be described with reference to FIG.

FIG. 15 is a flow chart showing an example of a specific processing procedure of negative number calculation processing. In the flowchart of FIG. 15, first, a period T from the disease name start date to the outcome date of the disease name information selected in step S1201 is specified (step S1501). Next, the information processing apparatus 101 extracts medical record information corresponding to the patient ID of the selected disease name information from the medical record DB 230 (step S1502).

Then, the information processing apparatus 101 selects unselected medical chart information from among the extracted medical chart information whose description date is included in the specified period T (step S1503). Next, the information processing apparatus 101 morphologically analyzes the article of the selected medical record information, and decomposes the article into word (morpheme) units (step S1504).

Then, the information processing apparatus 101 extracts words of a specific part of speech from the decomposed words (morphemes) (step S1505). Next, the information processing apparatus 101 calculates the number of appearances of each extracted word in the article of the selected medical record information (step S1506).

Then, the information processing apparatus 101 adds the calculated appearance count of each word to the number of negatives of each word in the negative-positive number table 800 corresponding to the disease name of the selected disease name information (step S1507). If there is no record of words corresponding to the disease name in the disease name information in the negative-positive number table 800, the information processing apparatus 101 newly creates a record of words corresponding to the disease name in the disease name information.

Next, the information processing apparatus 101 determines whether or not there is unselected medical chart information that has not been selected from the medical chart information that includes the entry date in the period T (step S1508). Here, if there is unselected medical chart information (step S1508: Yes), the information processing apparatus 101 returns to step S1503.

On the other hand, if there is no unselected medical record information (step S1508: No), the information processing apparatus 101 returns to the step that called the negative number calculation process. As a result, it is possible to count the number of occurrences of words in articles of medical record information classified as negative (medical record information with the outcome of "death").

Next, a specific processing procedure of the number of positives calculation processing in step S1207 shown in FIG. 12 will be described with reference to FIG.

FIG. 16 is a flow chart showing an example of a specific processing procedure of the number-of-positives calculation process. In the flowchart of FIG. 16, first, a period T from the disease name start date to the outcome date of the disease name information selected in step S1201 is specified (step S1601). Next, the information processing apparatus 101 extracts medical record information corresponding to the patient ID of the selected disease name information from the medical record DB 230 (step S1602).

Then, the information processing apparatus 101 selects unselected medical chart information from among the extracted medical chart information whose entry date is included in the specified period T (step S1603). Next, the information processing apparatus 101 morphologically analyzes the article of the selected medical record information, and decomposes the article into word (morpheme) units (step S1604).

Then, the information processing apparatus 101 extracts words of a specific part of speech from the decomposed words (morphemes) (step S1605). Next, the information processing apparatus 101 calculates the number of appearances of each extracted word in the article of the selected medical record information (step S1606).

Then, the information processing apparatus 101 adds the calculated number of appearances of each word to the positive number of each word in the negative-positive number table 800 corresponding to the disease name of the selected disease name information (step S1607).

Next, the information processing apparatus 101 determines whether or not there is unselected medical chart information that has not been selected from among the medical chart information whose description date is included in the period T (step S1608). Here, if there is unselected medical record information (step S1608: Yes), the information processing apparatus 101 returns to step S1603.

On the other hand, if there is no unselected medical chart information (step S1608: No), the information processing apparatus 101 returns to the step that called the positive number calculation process. This makes it possible to count the number of occurrences of words in articles of medical record information classified positively (medical record information whose outcome is “remission or cure”).

(Document search processing procedure of information processing apparatus 101)
Next, the document search processing procedure of the information processing apparatus 101 will be described with reference to FIGS. 17 and 18. FIG.

17 and 18 are flowcharts showing an example of the document search processing procedure of the information processing apparatus 101. FIG. In the flowchart of FIG. 17, first, the information processing apparatus 101 determines whether or not a search character string has been received from the client apparatus 201 (step S1701).

Here, the information processing apparatus 101 waits for reception of the search character string (step S1701: No). Then, when the search character string is received (step S1701: Yes), the information processing apparatus 101 identifies the disease name of the patient to be treated (step S1702). The information specifying the disease name of the patient to be treated may be associated with the search character string, or may be inquired of the client device 201 or the electronic medical record system.

Next, the information processing apparatus 101 morphologically analyzes the search character string to decompose the search character string into word (morpheme) units (step S1703). Then, the information processing apparatus 101 extracts words of a specific part of speech from the decomposed words (morphemes) (step S1704). Next, the information processing apparatus 101 refers to the negative-positive DB 240 for each extracted word, and specifies the negative-positive frequency corresponding to the disease name of the specified patient to be treated (step S1705).

Then, the information processing apparatus 101 calculates the negative-positive frequency of the search character string by accumulating the negative-positive frequency of each specified word (step S1706). Next, the information processing apparatus 101 searches for an article of medical record information corresponding to the search character string from the medical record DB 230 (step S1707), and proceeds to step S1801 shown in FIG.

Note that in step S1707, if the article of the chart information is not searched, the information processing apparatus 101 outputs a search result indicating that the article corresponding to the search character string was not searched, for example. A series of processes according to the flowchart is ended.

In the flowchart of FIG. 18, first, the information processing apparatus 101 selects an unselected article from among articles of one or a plurality of retrieved medical record information (step S1801). Next, the information processing apparatus 101 morphologically analyzes the selected article and decomposes the article into words (morphemes) (step S1802).

Then, the information processing apparatus 101 extracts words of a specific part of speech from the decomposed words (morphemes) (step S1803). Next, the information processing apparatus 101 refers to the negative-positive DB 240 for each extracted word, and specifies the negative-positive frequency corresponding to the disease name of the specified patient to be treated (step S1804).

Then, the information processing apparatus 101 calculates the negative-positive frequency of the selected article by dividing the value obtained by integrating the negative-positive frequency of each specified word by the number of extracted words (step S1805). Next, the information processing apparatus 101 determines whether or not there is an unselected article that has not been selected among the retrieved articles of the one or more pieces of medical chart information (step S1806).

Here, if there is an unselected article (step S1806: Yes), the information processing apparatus 101 returns to step S1801. On the other hand, if there is no unselected article (step S1806: No), the information processing apparatus 101 determines the priority of each article based on the calculated negative-positive frequency of each article and the negative-positive frequency of the search character string ( step S1807).

Then, the information processing apparatus 101 outputs to the client apparatus 201 a search result in which one or more articles of medical record information that have been searched are arranged in descending order of the determined priority (step S1808), and a series of processing according to this flowchart is performed. exit. As a result, it is possible to output search results that take into account positive or negative conditions according to the search character string.

In step S1807, the information processing apparatus 101 may further determine the display mode of each article based on the negative-positive frequency of each article. In this case, the information processing apparatus 101 outputs each article in the determined display mode in step S1808.

As described above, according to the information processing apparatus 101 according to the embodiment, referring to the first storage unit 710 in which documents that can be classified as positive or negative are registered, a word to be processed is classified as positive or negative. It is possible to calculate the negative-positive frequency of a word based on the appearance frequency of the word in each negatively classifiable document, and register the calculated negative-positive frequency in the second storage unit 720 in association with the word. The negative-positive frequency of a word is an index value indicating the tendency of the word to have positive or negative connotations.

This makes it possible to identify the tendency of positive or negative connotations of words, and furthermore realizes document retrieval taking into account positive or negative conditions.

Further, according to the information processing apparatus 101, when a search character string is received, documents corresponding to the search character string are searched from the document group, and one or more documents corresponding to the search character string are searched, the second By referring to the storage unit 720, the negative-positive frequency of each document can be calculated based on the negative-positive frequency corresponding to the words contained in each document. Further, according to the information processing apparatus 101, it is possible to refer to the second storage unit 720 and calculate the negative-positive frequency of the search character string based on the negative-positive frequency corresponding to the word included in the search character string. Then, according to the information processing apparatus 101, it is possible to output the search result corresponding to the search character string based on the calculated negative-positive frequency of each document and the negative-positive frequency of the search character string. The document group is, for example, a document group stored in the first storage unit 710 .

As a result, it is possible to estimate the tendency of positive or negative connotations of documents containing the words from the tendency of positive or negative connotations of words, and perform document retrieval with positive or negative conditions added.

Further, according to the information processing apparatus 101, based on the negative-positive frequency of each document and the negative-positive frequency of the search character string, priority is given to documents having a similar negative-positive frequency to the search character string among one or more retrieved documents. Can be output as search results.

As a result, it is possible to preferentially output documents (for example, medical record articles) that tend to have positive or negative connotations at the same level as the search character string as search results.

Further, according to the information processing apparatus 101, based on the negative-positive frequency of the search character string, the search character string is classified into positive or negative, and when the search character string is classified as positive, based on the negative-positive frequency of each document, Among the retrieved one or more documents, a document that tends to have a strong positive connotation can be preferentially output. Further, according to the information processing apparatus 101, when the retrieval character string is classified as negative, priority is given to documents having a strong negative connotation among one or more retrieved documents based on the negative-positive frequency of each document. can be output as

This makes it possible to determine whether documents with a strong positive or negative tendency are preferentially output according to whether the search character string is classified as positive or negative.

Further, according to the information processing apparatus 101, the display mode of each document is determined based on the negative positive frequency of each document, and each document is output in the determined display mode when outputting each document as a search result. can be done.

As a result, when displaying search results corresponding to a search character string, it is possible to distinguish between documents with positive connotations and documents with negative connotations.

Further, according to the information processing apparatus 101, articles of patient's medical chart information, which are documents stored in the first storage unit 710 (medical chart DB 230), can be classified into positive or negative according to the outcome of the patient. can.

This allows, for example, classifying medical record entries written during treatment of an illness as positive or negative depending on whether the patient's outcome is "remission or cure" or "death". can.

Further, according to the information processing apparatus 101, out of the documents (articles of medical record information) stored in the first storage unit 710 (medical record DB 230), word can be calculated. Then, according to the information processing apparatus 101, the calculated negative-positive frequency can be registered in the second storage unit 720 (negative-positive DB 240) in association with the word and each disease name.

As a result, the negative-positive frequency of words used in medical record articles can be obtained by considering how the words are used for each disease name. For example, a word that is used with a negative connotation in one disease may be used with a positive connotation in another disease. By considering the usage characteristics of each disease name, it is possible to accurately identify the tendency of positive or negative connotations of words.

Further, according to the information processing apparatus 101, the second storage unit 720 (negative/positive DB 240) is referred to, and words included in each retrieved document (article of medical record information) correspond to the disease name of the patient to be treated. The negative-positive frequency of each document can be calculated based on the negative-positive frequency of words contained in each document. Further, according to the information processing apparatus 101, the second storage unit 720 is referred to, the negative-positive frequency corresponding to the disease name of the patient to be treated is specified for the word included in the search character string, and the specified search character string It is possible to calculate the negative-positive frequency of the search character string based on the negative-positive frequency of words included in .

This makes it possible to determine the frequency of negative positives in search strings and medical record articles by taking into consideration the usage of words for each disease name, and accurately identify trends in positive or negative connotations of search strings and medical chart articles. It becomes possible to

For these reasons, according to the information processing apparatus 101 and the document search system 200 according to the embodiment, by realizing a document search with positive or negative conditions added, the user can quickly and easily find necessary information. In this way, it is possible to improve work efficiency when searching for documents.

The document retrieval method described in this embodiment can be realized by executing a prepared program on a computer such as a personal computer or a workstation. The document search program is recorded in a computer-readable recording medium such as a hard disk, flexible disk, CD-ROM, DVD, USB memory, etc., and executed by being read from the recording medium by a computer. Also, the document search program may be distributed via a network such as the Internet.

Further, the information processing apparatus 101 described in the present embodiment can also be realized by application specific ICs such as standard cells and structured ASICs (Application Specific Integrated Circuits), and PLDs (Programmable Logic Devices) such as FPGAs.

Further, the following additional remarks are disclosed with respect to the above-described embodiment.

(Appendix 1) Referring to a first storage unit in which documents that can be classified as positive or negative are registered, a word to be processed is based on the appearance frequency of the word in each of the documents that can be classified as positive or negative. , calculating an index value indicating the tendency of the positive or negative connotation of the word,
registering the calculated index value in a second storage unit in association with the word;
A document search program characterized by causing a computer to execute processing.

(Appendix 2) Receiving a search string,
searching for a document corresponding to the search string from the group of documents;
when one or more documents corresponding to the search character string are retrieved, referring to the second storage unit, based on index values corresponding to words contained in each of the one or more documents, calculating an index value for each of the documents;
referring to the second storage unit and calculating the index value of the search character string based on the index value corresponding to the word included in the search character string;
outputting a search result corresponding to the search character string based on the calculated index value of each document and the index value of the search character string;
The document search program according to appendix 1, which causes the computer to execute processing.

(Appendix 3) The process of outputting
Based on the index value of each document and the index value of the search character string, among the one or more documents, a document having an index value close to the search character string is preferentially output as a search result. A document search program according to appendix 2, characterized by:

(Appendix 4) causing the computer to perform a process of classifying the search string as positive or negative based on the index value of the search string;
The output process is
when the search character string is positively classified, based on the index value of each document, priority is given to a document having a strong positive connotation among the one or more documents, and
when the search character string is classified as negative, preferentially outputs a document having a strong negative connotation among the one or more documents based on the index value of each document;
The document search program according to appendix 2, characterized by:

(Appendix 5) causing the computer to execute a process of determining a display mode of each document based on the index value of each document;
The output process is
5. The document retrieval program according to any one of attachments 2 to 4, characterized in that, when outputting each document as a retrieval result, each document is output in the determined display mode.

(Appendix 6) Documents stored in the first storage unit are patient chart information, which can be classified as positive or negative according to the outcome of the patient. A document retrieval program according to any one of the above.

(Appendix 7) The document stored in the first storage unit is associated with the disease name of the patient,
The process of calculating the index value of the word includes:
calculating an index value indicating the tendency of the word to have a positive or negative connotation based on the appearance frequency of the word in each document corresponding to each disease name among the documents stored in the first storage unit;
The process of registering
7. The document search program according to appendix 6, wherein the calculated index value is associated with the word and each disease name and registered in the second storage unit.

(Appendix 8) The process of calculating the index value of each document is
identifying an index value corresponding to a disease name of a patient to be treated for a word contained in each of the retrieved documents by referring to the second storage unit;
calculating an index value of each of the documents based on the index values of the words contained in each of the identified documents;
The process of calculating the index value of the search character string includes:
referring to the second storage unit to specify an index value corresponding to the disease name of the patient to be treated for the word included in the search character string;
The document search program according to appendix 7, wherein the index value of the search character string is calculated based on the index value of the specified word included in the search character string.

(Appendix 9) The document search program according to any one of Appendices 2 to 8, wherein the document group is a document group stored in the first storage unit.

(Appendix 10) The document search program according to any one of Appendices 1 to 9, wherein the words to be processed are words included in documents stored in the first storage unit. .

(Appendix 11) Referring to a first storage unit in which documents that can be classified as positive or negative are registered, a word to be processed is based on the appearance frequency of the word in each of the documents that can be classified as positive or negative. , calculating an index value indicating the tendency of the positive or negative connotation of the word,
registering the calculated index value in a second storage unit in association with the word;
A document retrieval method characterized in that processing is executed by a computer.

(Appendix 12) Referring to a first storage unit in which documents that can be classified as positive or negative are registered, a word to be processed is based on the appearance frequency of the word in each of the documents that can be classified as positive or negative. , a calculation unit that calculates an index value indicating the tendency of the word to have a positive or negative connotation;
a registration unit that associates the index value calculated by the calculation unit with the word and registers the index value in a second storage unit;
A document retrieval system comprising:

101 information processing device 110,710 first storage unit 120,720 second storage unit 111,112 document 200 document search system 201 client device 210 network 220 disease name DB
230 medical record database
240 negative positive DB
300 Bus 301 CPU
302 memory 303 disk drive 304 disk 305 communication I/F
306 portable recording medium I/F
307 portable recording medium 701 first reception unit 702 calculation unit 703 registration unit 704 second reception unit 705 search unit 706 determination unit 707 output unit 800 negative/positive number table 910, 1100 keyword search screen 920, 1000 search result screen 921, 922, 923, 924 Medical records 931, 932, 933, 934 Negative and positive frequency information

Claims (9)

By referring to a first storage unit in which positively classified documents and negatively classified documents are registered, words included in at least one of the positively classified documents and the negatively classified documents calculating an index value indicating the tendency of the word to have a positive or negative connotation based on the appearance frequency of the word in each of the positively classified documents and the negatively classified documents;
registering the calculated index value in a second storage unit in association with the word;
accepts a search string,
searching for a document corresponding to the search string from the group of documents;
when one or more documents corresponding to the search character string are retrieved, referring to the second storage unit, based on index values corresponding to words contained in each of the one or more documents, calculating an index value for each of the documents;
referring to the second storage unit and calculating the index value of the search character string based on the index value corresponding to the word included in the search character string;
outputting a search result corresponding to the search character string based on the calculated index value of each document and the index value of the search character string;
A document search program characterized by causing a computer to execute processing. The output process is
Based on the index value of each document and the index value of the search character string, among the one or more documents, a document having an index value close to the search character string is preferentially output as a search result. 2. A document search program according to claim 1. causing the computer to perform a process of classifying the search string as positive or negative based on the index value of the search string;
The output process is
when the search character string is positively classified, based on the index value of each document, priority is given to a document having a strong positive connotation among the one or more documents, and
when the search character string is classified as negative, preferentially outputs a document having a strong negative connotation among the one or more documents based on the index value of each document;
2. The document search program according to claim 1, characterized by: causing the computer to execute a process of determining a display mode of each of the documents based on the index value of each of the documents;
The output process is
4. The document retrieval program according to any one of claims 1 to 3, wherein when outputting each document as a retrieval result, each document is output in the determined display mode. 5. The document stored in the first storage unit is patient medical chart information, and is classified into positive or negative according to the outcome of the patient. the document retrieval program described in . The document stored in the first storage unit is associated with the disease name of the patient,
The process of calculating the index value of the word includes:
calculating an index value indicating the tendency of the word to have a positive or negative connotation based on the appearance frequency of the word in each document corresponding to each disease name among the documents stored in the first storage unit;
The process of registering
6. The document retrieval program according to claim 5, wherein the calculated index value is associated with the word and each disease name and registered in the second storage unit. The process of calculating the index value of each document includes:
identifying an index value corresponding to a disease name of a patient to be treated for a word contained in each of the retrieved documents by referring to the second storage unit;
calculating an index value of each of the documents based on the index values of the words contained in each of the identified documents;
The process of calculating the index value of the search character string includes:
referring to the second storage unit to specify an index value corresponding to the disease name of the patient to be treated for the word included in the search character string;
7. The document search program according to claim 6, wherein an index value of said search character string is calculated based on an index value of a word included in said specified search character string. By referring to a first storage unit in which positively classified documents and negatively classified documents are registered, words included in at least one of the positively classified documents and the negatively classified documents calculating an index value indicating the tendency of the word to have a positive or negative connotation based on the appearance frequency of the word in each of the positively classified documents and the negatively classified documents;
registering the calculated index value in a second storage unit in association with the word;
accepts a search string,
searching for a document corresponding to the search string from the group of documents;
when one or more documents corresponding to the search character string are retrieved, referring to the second storage unit, based on index values corresponding to words contained in each of the one or more documents, calculating an index value for each of the documents;
referring to the second storage unit and calculating the index value of the search character string based on the index value corresponding to the word included in the search character string;
outputting a search result corresponding to the search character string based on the calculated index value of each document and the index value of the search character string;
A document retrieval method characterized in that processing is executed by a computer. By referring to a first storage unit in which positively classified documents and negatively classified documents are registered, words included in at least one of the positively classified documents and the negatively classified documents a calculation unit that calculates an index value indicating a tendency of the positive or negative connotation of the word based on the appearance frequency of the word in each of the positively classified documents and the negatively classified documents;
a registration unit that associates the index value calculated by the calculation unit with the word and registers the index value in a second storage unit;
a reception unit that receives a search string;
a search unit that searches for a document corresponding to the search character string received by the reception unit from a group of documents;
index corresponding to words contained in each of the one or more documents by referring to the second storage unit when one or more documents corresponding to the search character string are retrieved by the search unit; calculating the index value of each document based on the value, referring to the second storage unit, and calculating the index value of the search character string based on the index value corresponding to the word included in the search character string a calculation unit for
an output unit configured to output a search result corresponding to the search character string based on the index value of each document and the index value of the search character string calculated by the calculation unit;
A document retrieval system comprising:

Images