Investigation support method and investigation support system

Abstract

The present invention provides an investigation support method and an investigation support system. This investigation support method includes: extracting a plurality of documents on the basis of a set investigation theme; allowing a user to review the plurality of documents sequentially; determining whether or not a document under review includes an unregistered noise word; storing the unregistered noise word in a storage device of a computer when input of the unregistered noise word is received; and, when the plurality of documents are reviewed, allowing the noise word stored in the storage device and words and phrases other than the noise word to be distinguishable from each other.

Description

Investigation Support Method and Investigation Support System

[0001] The present invention relates to an investigation support method and an investigation support system.

[0002] In literature research such as patents and papers, relevant documents are extracted by search from a database and their contents are evaluated. However, in many cases, a large number of irrelevant documents are mixed in the extracted documents. Therefore, there is a problem that a lot of time and labor are wasted in the investigation.

[0003] To address this problem, in recent years, a patent investigation support method using machine learning has been proposed (Patent Document 1). In this method, for unevaluated case data, the search target probability for a search target case and the noise target probability for a noise target case are obtained by machine learning (artificial intelligence), and based on the search target probability and the noise target probability of the unevaluated case data, it is plotted on a graph with the first axis being the search probability and the second axis being the noise probability, and the evaluation target case data can be selected after grasping the scatter situation of the plots from this graph.

[0004] Further, Patent Document 1 describes that at least one of the search conditions for a search target case and the noise conditions for a noise target case is input into an input field as a search condition, and a graph in which unevaluated case data that matches the search condition is plotted is created from the unevaluated case data plotted on the graph according to the search condition input into the input field.

[0005] Note that in addition to keywords, the search conditions can include technical classifications (for example, IPC, FI, F-term) and identification numbers assigned to cases (for example, application number, publication number, registration number).

[0006] Japanese Patent Application Laid-Open No. 2020-095303

[0007] Incidentally, FTO searches (Freedom to Operate searches) are a well-known method of patent searching. An FTO search is a search to confirm whether a company's technology can be implemented without infringing on the patent rights of other companies. A characteristic of FTO searches is that it is necessary to search for patents related to the target product, resulting in a large number of results. On the other hand, depending on the technical challenge tackled by the technology developer, there are cases where thousands of other companies' patents extracted by the search formula end up being noise (irrelevant documents). In other words, even if there are no problematic other companies' patents, the screening officer (the person in charge of selecting relevant documents) needs to thoroughly investigate the thousands of patents extracted by the search formula. Under these circumstances, the screening officer sometimes felt stressed about not overlooking even one problematic patent, and became exhausted from the work.

[0008] Furthermore, high levels of noise negatively impact machine learning. Specifically, overfitting due to large amounts of noisy data (noise overfitting) creates a new problem where machine learning models overlook problematic patents.

[0009] Furthermore, while the patent search support method described in Patent Document 1 makes it possible to appropriately narrow down the data of unevaluated cases that require review and scrutiny from a large number of unevaluated case data based on a set theme, Patent Document 1 does not describe how to review a large number of unevaluated case data, nor does it describe how to improve the speed of the review process.

[0010] Furthermore, in Patent Document 1, the noise conditions such as keywords entered in the input field are one of the search conditions for searching for unevaluated case data that matches the noise conditions from among the unevaluated case data, and are not used for peer review. Also, if the number of keywords set as noise conditions increases, the number of unevaluated case data that match those noise conditions will decrease.

[0011] This invention has been made in view of these circumstances, and aims to provide a research support method and a research support system that enable efficient searching of related documents in a search of multiple documents, such as patent documents and technical documents.

[0012] The invention according to the first aspect is a research support method that extracts multiple documents based on a set research theme, allows the user to sequentially review the multiple documents, determines whether the documents contain unregistered noise words, stores the unregistered noise words in the computer's memory when input of unregistered noise words is received, and makes it possible to distinguish between the noise words stored in the memory and other words when multiple documents are reviewed.

[0013] According to the first embodiment, noise words and other words can be easily recognized, and the review time for each document reviewed sequentially based on the recognized noise words can be shortened. In addition, while reviewing a document, the user can find unregistered noise words that contribute to the determination of whether or not the document is subject to further review, and these unregistered noise words can be sequentially stored in the computer's memory (additional registration). This makes it possible to optimize the noise words stored in the memory.

[0014] In the second embodiment of the research support method, when a user sequentially reviews multiple documents, it is preferable to display the multiple documents sequentially on the display screen, and to highlight the noise words stored in the storage device on the display screen for the multiple documents displayed sequentially on the display screen.

[0015] According to the second embodiment, the user can easily see on the display screen whether or not the document contains noise words, and if so, those noise words.

[0016] The third aspect of the research support method, in the first aspect, allows the user to sequentially review multiple documents, automatically reading them aloud in sequence, and to switch the automatic reading method between noise words stored in the memory device and words other than noise words for the multiple documents that are read aloud in sequence.

[0017] According to the third aspect, the user can easily determine whether or not the automatically read-aloud document contains noise words, and if so, distinguish those noise words.

[0018] The investigation support method relating to the fourth aspect is such that, in any of the first to third aspects, the multiple documents are multiple patent documents.

[0019] According to the fourth embodiment, the time required for reviewing patent documents can be reduced.

[0020] In the fifth aspect of the investigation support method, when noise words and phrases other than noise words are distinguishable in the fourth aspect, the noise words and phrases other than noise words are distinguishable at least within the claims of the patent document.

[0021] According to the fifth aspect, since the relationship between a patent document and a research topic is often based on the description of the claims in the patent document, it is effective to make it possible to distinguish between noise words and other words in the claims when determining the relationship. In addition, noise included in the specification of the patent document may also be highlighted.

[0022] The sixth aspect of the search support method makes it possible to distinguish between a first patent document in which the noise word is not included in the claims and a second patent document in which the noise word is included in the claims, in the fourth aspect.

[0023] According to the sixth aspect, it is possible to distinguish between patent documents that do not contain noise words in their claims (first patent document) and patent documents that do contain noise words in their claims (second patent document) without peer review.

[0024] The seventh aspect of the search support method, in the fourth aspect, involves assigning identification information to at least one of the first patent document and the second patent document to distinguish between a first patent document in which the noise word is not included in the claims and a second patent document in which the noise word is included in the claims, wherein the identification information is information that can be used when rearranging the first patent document and the second patent document in multiple patent documents, or when selecting only the first patent document from multiple patent documents as a patent document for peer review.

[0025] According to the seventh embodiment, for example, by rearranging the first patent document to rank higher than the second patent document based on identification information, or by selecting only the first patent document as the patent document for review, the review of the second patent document can be omitted, thereby shortening the review time for the patent document search.

[0026] The investigation support method for the eighth aspect involves receiving the noise word to be memorized from the input reception field in any of the first to seventh aspects.

[0027] According to the eighth aspect, users can easily input noise words by using the input reception field.

[0028] The survey support method according to the ninth embodiment involves displaying the input reception field on a display screen that sequentially displays multiple documents, as described in the eighth embodiment.

[0029] According to the ninth aspect, the user can input noise words using the input field displayed on the display screen.

[0030] The investigation support method relating to the 10th aspect accepts the editing of noise words in the 8th or 9th aspect.

[0031] According to the tenth embodiment, by accepting editing of noise words, it becomes possible to add, modify, or delete noise words, and the noise words stored in the memory can be optimized, resulting in more accurate correlation between literature and research themes.

[0032] The investigation support method relating to the 11th embodiment involves displaying a prompt on the display screen to enter noise words, as in the 9th or 10th embodiment.

[0033] According to the eleventh aspect, users can be encouraged to use noise words in their peer review.

[0034] In the investigation support method relating to the 12th aspect, in any of the 1st to 11th aspects, the noise words to be stored are received based on user operations on the operation unit.

[0035] The investigation support method according to the 13th embodiment is, in the 12th embodiment, a user operation in the operation unit which is an input operation and a registration operation of unregistered noise words performed using a display screen that displays multiple documents in sequence.

[0036] The investigation support method according to the 14th embodiment, in the 13th embodiment, includes a cursor operation in which the input operation selects a noise word on the display screen, and a registration operation which is an instruction operation to store the noise word selected by the cursor operation in the memory device.

[0037] In the 14th embodiment, the investigation support method is such that cursor operation includes at least one of mouse operation, eye-tracking operation, gesture operation, and touchpad operation, and instruction operation includes at least one of mouse click operation, touchpad tap operation, keyboard input, voice input, eye-tracking input, gesture input, and electroencephalogram input.

[0038] According to the 15th embodiment, the user can select a desired noise word on the display screen by cursor operation such as mouse operation, and can store the selected noise word in the memory device by registration operation such as mouse click operation. This makes it possible to easily store the desired noise word in the memory device (and thereby make it possible to distinguish the noise word from other words).

[0039] The investigation support method relating to the 16th aspect is such that, in any of the 13th to 15th aspects, the display screen is provided with an input reception field into which the noise words to be memorized are entered, and the input operation is to input the noise words into the input reception field using a keyboard, voice input, or by copying and pasting or dragging and dropping them into the input reception field.

[0040] According to the 16th embodiment, noise words are accepted by keyboard input or voice input into the input input field displayed on the display screen, or by copying and pasting or dragging and dropping noise words into the input input field displayed on the display screen.

[0041] In the investigation support method according to the 17th aspect, in any of the 1st to 7th aspects, the noise words to be memorized are received based on voice input.

[0042] In the investigation support method according to the 18th aspect, in any of the 1st to 17th aspects, it is determined whether there is an expression requiring attention different from the noise words in the document, and when there is an expression requiring attention, the expression requiring attention and the words other than the expression requiring attention are made distinguishable.

[0043] According to the 18th aspect, since the user can distinguish between an expression requiring attention and words other than the expression requiring attention, which affect the determination of whether there is a relevance to the investigation theme, the user can easily recognize the expression requiring attention from the document, and it becomes possible to determine the relevance considering the recognized expression requiring attention.

[0044] In the investigation support method according to the 19th aspect, in any of the 1st to 17th aspects, it is determined whether there is an expression requiring attention different from the noise words in the document, and when there is an expression requiring attention in the vicinity of the noise words, the distinguishability between the noise words and the words other than the noise words is reduced.

[0045] When there is an expression requiring attention in the vicinity of the noise words, there may be a case where it can be regarded as not being a noise word due to the expression requiring attention, and it is preferable to reduce the distinguishability between the noise words and the words other than the noise words (not making the noise words prominent).

[0046] In the investigation support method according to the 20th aspect, in the 18th or 19th aspect, the expression requiring attention includes at least any one of alternative descriptions, descriptions related to conditions, conditional branches, and negative expressions.

[0047] According to the 20th aspect, the user can easily recognize at least any one of alternative descriptions, conditional branches, and negative expressions, which are expressions requiring attention, from the document.

[0048] In the investigation support method according to the 21st aspect, in any of the 1st to 20th aspects, when the user sequentially examines a plurality of documents, it is received whether the examination of the document has been performed, and for each document, at least a first bookmark indicating whether the examination has been performed is attached.

[0049] According to the 21st embodiment, each document is given a first bookmark indicating whether or not it has undergone peer review. By checking whether or not the first bookmark is given, it is possible to prevent multiple documents from being overlooked during peer review.

[0050] The research support method according to the 22nd embodiment, in the 21st embodiment, determines the relationship between multiple documents and the research theme when the user sequentially reviews the documents, and provides each document with a second bookmark indicating the result of the relationship determination.

[0051] According to the 22nd embodiment, a second bookmark indicating the determination of relevance to the research theme is attached to each document. This allows researchers to narrow down the documents requiring further examination based on the second bookmark after the peer review of multiple documents is completed, enabling them to carry out the time-consuming examination.

[0052] The research support method according to the 23rd embodiment, in the 22nd embodiment, determines the relationship between the literature and the research theme if the literature does not contain noise words, and determines the relationship between the literature and the research theme by referring to the noise words if the literature does contain noise words.

[0053] According to the 23rd embodiment, if the literature does not contain noise words, the relationship between the literature and the research theme can be determined (by performing a normal determination without referring to noise words), and if the literature contains noise words, the relationship between the literature and the research theme can be determined by referring to the noise words.

[0054] The research support method according to the 24th embodiment determines, in the 22nd embodiment, whether or not there are expressions of concern different from noise words in the literature; if the literature does not contain noise words, it determines the relationship between the literature and the research theme; if the literature contains noise words, it determines the relationship between the literature and the research theme by referring to the noise words; and if the literature contains both noise words and expressions of concern, it determines the relationship between the literature and the research theme by referring to the noise words and expressions of concern.

[0055] According to the 24th aspect, the user can determine the relevance between the literature and the research theme if the literature does not contain noise words, determine the relevance between the literature and the research theme by referring to the noise words if the literature contains noise words, and determine the relevance between the literature and the research theme by referring to the noise words and the expression of concern if the literature contains both noise words and expression of concern.

[0056] The invention according to the 25th aspect is a research support system equipped with a processor, wherein the processor extracts multiple documents based on a set research theme, and when the user sequentially reviews the multiple documents, if the document being reviewed contains an unregistered noise word, the processor accepts the unregistered noise word based on user operation on the operation unit, stores the accepted noise word in the computer's storage device, and when multiple documents are reviewed, the system makes it possible to distinguish between the noise word stored in the storage device and other words.

[0057] In the 26th aspect of the research support system, in the 25th aspect, the processor displays multiple documents sequentially on a display screen when a user sequentially reviews multiple documents, and highlights noise words stored in a memory device on the display screen for the multiple documents displayed sequentially on the display screen.

[0058] In the 27th aspect of the research support system, in the 25th aspect, the processor automatically reads aloud multiple documents in order when a user sequentially reviews multiple documents, and switches the automatic reading method for noise words stored in the memory and words other than noise words for the multiple documents that are read aloud in order.

[0059] In the investigation support system according to the 28th aspect, in any of the 25th to 27th aspects, the processor extracts multiple patent documents as multiple documents.

[0060] In the 29th aspect of the investigation support system, in the 28th aspect, the processor makes it possible to distinguish between noise words and phrases other than noise words, at least within the claims of the patent document.

[0061] In the 30th embodiment, the research support system, in the 28th or 29th embodiment, enables the processor to distinguish between a first patent document in which the noise word is not included in the claims and a second patent document in which the noise word is included in the claims.

[0062] The search support system according to the 31st embodiment, in embodiments 28 to 30, provides a processor that assigns identification information to at least one of a first patent document in which the noise word is not included in the claims and a second patent document in which the noise word is included in the claims, wherein the identification information is information that can be used when rearranging the first patent document and the second patent document in multiple patent documents, or when selecting only the first patent document from multiple patent documents as a patent document for peer review.

[0063] In the 32nd aspect of the research support system, in any of the 25th to 31st aspects, the processor displays an input field for inputting noise words on a display screen that sequentially displays multiple documents, and accepts noise words to be stored from the input field.

[0064] In the third aspect of the investigation support system, the processor accepts editing of noise words.

[0065] In the investigation support system relating to the 34th aspect, in the 32nd or 33rd aspect, the processor causes a description prompting the input of noise words to be displayed on the display screen.

[0066] The research support system according to the 35th embodiment, in any of the 25th to 34th embodiments, has an input reception field on the display screen that sequentially displays multiple documents, and the operation unit includes a keyboard, and the keyboard is assigned a first shortcut key that enables the input of noise words into the input reception field, and when the first shortcut key is operated, the cursor on the display screen is moved to a predetermined position in the input reception field.

[0067] According to the 35th embodiment, a first shortcut key is assigned to the keyboard that moves the cursor on the display screen to a predetermined position in the noise word setting field. Therefore, by simply pressing the first shortcut key once, the cursor (pointer) for character input can be moved to a predetermined position in the noise word setting field (the starting position where unregistered noise words can be entered), and the cursor for character input can be moved in a short amount of time.

[0068] In the investigation support system relating to the 36th aspect, in any of the 32nd to 35th aspects, the processor accepts noise words to be stored based on keyboard input into an input reception field, voice input, or copy and paste or drag and drop.

[0069] In the investigation support system according to the 37th aspect, in any of the 25th to 31st aspects, the processor receives noise words to be stored based on voice input.

[0070] The investigation support system according to the 38th embodiment, in the 35th or 36th embodiment, is provided on the display screen with a first soft button for registration that stores noise words entered in the input reception field in a storage device, and the operation unit includes a keyboard, on which a second shortcut key is assigned to move the cursor over the first soft button.

[0071] According to the 38th embodiment, a second shortcut key is assigned to the keyboard to move the cursor to the first soft button for registration. Therefore, after a noise word is entered in the noise word setting field, the cursor can be moved to the first soft button by simply pressing the second shortcut key once, and the first soft button can be operated immediately (the operation to register an unregistered noise word can be performed immediately).

[0072] In the investigation support system according to the 39th embodiment, in any of the 25th to 38th embodiments, the processor determines whether or not a noteworthy expression different from noise words exists in the literature, and if a noteworthy expression exists, it makes it possible to distinguish the noteworthy expression from other words or phrases.

[0073] In the investigation support system according to the 40th embodiment, in any of the 25th to 39th embodiments, the processor determines whether or not a noteworthy expression different from a noise word exists in the literature, and if a noteworthy expression exists near a noise word, it reduces the indistinguishability between the noise word and other words.

[0074] In the 41st embodiment, the investigation support system, in the 39th or 40th embodiment, includes at least one of the following as a cautionary expression: an alternative statement, a statement relating to a condition, a conditional branch, and a negative expression.

[0075] In the 42nd aspect of the research support system, in any of the 25th to 41st aspects, the processor receives information on whether or not a peer review has been conducted on the document, and assigns a first bookmark to each document indicating at least whether or not a peer review has been conducted.

[0076] In the 43rd embodiment of the research support system, in the 42nd embodiment, a display screen that sequentially displays multiple documents is provided with a second soft button for receiving input indicating the completion of peer review, and the processor receives whether or not peer review has been performed based on the user operation of the second soft button.

[0077] According to the 43rd embodiment, when a second soft button provided on the display screen is operated, a first bookmark indicating that a peer review has been performed can be added to the document displayed on the display screen.

[0078] In the 44th aspect of the investigation support system, in the 42nd or 43rd aspect, if the document contains an unregistered noise word and the unregistered noise word is stored in the memory, the processor acknowledges that a peer review has been conducted on the document.

[0079] According to the 44th embodiment, if a document contains an unregistered noise word and the unregistered noise word is stored in a memory device, a first bookmark indicating that the document has undergone peer review can be attached to the document in conjunction with the storage of the unregistered noise word.

[0080] In the 45th aspect of the research support system, in any of the 42nd to 44th aspects, the display screen that sequentially displays multiple documents is provided with two or more third soft buttons that indicate whether or not there is a relationship between the document and the research theme, or the degree of that relationship. The processor receives a determination result indicating whether or not there is a relationship or the degree of that relationship based on the user operation of the third soft buttons, and assigns a second bookmark to each document indicating the determination result of the relationship.

[0081] According to the 45th embodiment, the display screen showing the peer-reviewed literature is provided with two or more third soft buttons, including whether or not the literature is related to the research theme, or the degree of relatedness. By operating the third soft buttons, the user can add a second bookmark to each literature, indicating the determination result of whether or not the literature is related to the research theme, or the degree of relatedness.

[0082] According to the present invention, it is possible to efficiently search multiple documents, such as patent documents and technical documents, and to reduce the time required for the search.

[0083] Figure 1 is a flowchart showing an example of the FTO investigation procedure. Figure 2 is a diagram showing an overview of the investigation support system. Figure 3 is a block diagram showing the schematic configuration of the server and terminal device hardware. Figure 4 is a block diagram showing the processing functions realized by the processor. Figure 5 is a flowchart showing the first embodiment of the investigation support method according to the present invention. Figure 6 is a diagram showing an example of a display screen shown in list mode. Figure 7 is a diagram showing an example of a display screen shown in browsing mode. Figure 8 is a diagram showing an example of a display screen shown in browsing mode. Figure 9 is a diagram showing an example of a display screen shown in browsing mode. Figure 10 is a diagram showing an example of a display screen shown in browsing mode. Figure 11 is a diagram showing an example of a display screen shown in list mode. Figure 12 is a diagram showing a part of a flowchart for explaining the first embodiment. Figure 13 is a diagram showing a part of a flowchart for explaining the second embodiment.

[0084] Preferred embodiments of the present invention will be described below with reference to the attached drawings.

[0085] [Overview of FTO Investigation] First, let's explain the procedure for an FTO investigation. Figure 1 is a flowchart showing the procedure for a typical FTO investigation. As shown in Figure 1, an FTO investigation consists of steps such as clarifying the subject of the investigation (Step S1), formulating a search strategy (Step S2), conducting a database search (Step S3), reviewing extracted documents (Step S4), scrutinizing related documents (Step S5), risk assessment (Step S6), and preparing a report (Step S7). The following explanation will be based on patent documents, but technical documents may also be explained.

[0086] In clarifying the scope of the investigation (Step S1), the products or methods to be investigated are identified, the scope of the investigation is determined, and the investigation period is set.

[0087] When identifying the product or method to be investigated (such as a manufacturing method or measurement method), the technical characteristics of the product, such as its function, structure, and materials, should be clearly defined. Similarly, the technical characteristics of the method, such as its processes, conditions, and materials, should also be clearly defined. The effects and benefits of the product and method should also be clarified. By clarifying these technical characteristics and effects, relevant patent or technical literature can be efficiently searched. Based on these technical characteristics and effects, the research theme is then determined.

[0088] When determining the scope of the survey, you should decide on a country or region. For example, this could be determined by considering the planned sales area for the product, the areas where competitors operate, and regions where there is a significant amount of technological advancement and published technical literature.

[0089] When setting a search period, the period for patent documents is determined by considering the patent's validity period (usually within 20 years from the filing date). However, if necessary, past patent documents can be included without considering the patent's validity period. Including past patents allows for the identification of technologies whose rights have expired. For technical documents, the period can be set based on the publication date of the document, depending on the purpose of the search.

[0090] In developing a search strategy (Step S2), you will select keywords, identify relevant patent classifications, and list competitors.

[0091] When selecting keywords, choose appropriate keywords that describe the technical characteristics of the product and method. Including synonyms and related terms will enable a more comprehensive search. If you have conducted similar research in the past, you may refer to the keywords used at that time.

[0092] When identifying relevant patent classifications, you can use IPC (International Patent Classification), F-terms, and CPC (Cooperative Patent Classification), among others. Using patent classifications allows you to comprehensively search for patent documents related to your research topic, improving search accuracy. Similarly, using classifications in technical documents can prevent missed searches and improve search accuracy.

[0093] When listing competitors or inventors, identify companies, research institutions, inventors, or authors active in the technology field related to your research topic. This information can be gathered online and reduces the risk of missing important patent and technical documents.

[0094] In the database search (step S3), at least one database—a commercial database, a database provided by a national patent office, and an in-house database—is used to perform a more comprehensive search based on the search strategy. Multiple documents (patent documents or technical documents) related to the research theme are extracted. Typically, several thousand documents, such as 2,000 to 3,000, are extracted. When extracting documents, the database may be searched based on the information organized in the search strategy planning (step S2), or if there is already an extracted list of external documents, the database may be searched using that list, or if there is already an extracted list of internal documents in the database, it may be directly referenced. In the case of an FTO search, the number of documents to be peer-reviewed may be reduced by excluding documents that have lost their rights after patenting and documents that lost the possibility of patenting before patenting.

[0095] In the peer review of extracted literature (step S4), multiple documents extracted in the database search (step S3) are reviewed. In the peer review of extracted literature (step S4), documents that are highly likely to be related to the research theme are selected from among the multiple documents extracted. In the peer review, the goal is to reduce the number of documents that require time-consuming scrutiny (this is called the "garbage removal" process) and to reduce the number of documents that should be scrutinized; in other words, efficient and thorough selection is important here. Here, garbage removal and scrutiny may be defined as, for example, the act of sorting whether a document should be scrutinized or not without specifying the reason for the judgment, while scrutiny may be the act of confirming or specifying whether a document is relevant, the basis for the judgment of relevance, or points to note.

[0096] However, screening personnel, especially during the process of removing unwanted patents, often felt stressed and exhausted by having to meticulously examine thousands of patents extracted using search formulas, under the pressure of not wanting to miss even a single problematic patent. In other words, while efficient extraction is desired, it has not been achieved in reality. In addition, the process of removing unwanted patents is too time-consuming, and the recording of the reasoning behind the decisions and the specific points of judgment, which takes even more time, is almost always omitted. This also creates the problem that reviewing and sharing the work content is virtually impossible. Therefore, the inventor diligently considered the review of extracted documents and arrived at the present invention.

[0097] In the review of related literature (Step S5), the wording of the claims in the literature extracted in the peer review of the extracted literature (Step S4) is examined in detail, and the similarity (relevance in a deeper sense) to the product and method of the research topic is considered. The correspondence between the product and method and the components described in the claims is confirmed. Careful consideration is given to whether the product and method may fall within the technical scope. In the case of technical literature, the correspondence with the product and method is confirmed based on the entire document.

[0098] In the case of patent documents, the expiration date and annuity payment status of the patent are checked to determine whether the rights are actually valid. The existence of opposition proceedings, invalidation trials, and lawsuits is also investigated to assess the stability of the rights.

[0099] In the risk assessment (Step S6), based on the analysis results from the detailed examination, the risk level (high, medium, and low, etc.) is carefully determined, taking into account the similarity and interpretation of the scope of rights. For patents with a high risk level, attention is paid, and countermeasures are considered as a priority.

[0100] In report preparation (Step S7), the investigation theme, investigation methodology, analysis results, and risk assessment are summarized. The entire investigation process (investigation scope, investigation period, database, and search formula, etc.) and results are clearly described, and the basis for the risk assessment is described. Avoidance measures (design changes, etc.) and countermeasures (invalidation investigation, license negotiation, etc.) are proposed as needed. The FTO investigation is completed. In some cases, multiple FTO investigations may be conducted for the same investigation theme with modified search strategies.

[0101] [Disclosure of the present invention] In the peer review of extracted documents (step S4), unreviewed patent documents are displayed on the display screen of the display device, and those patent documents are reviewed. When the review of the patent documents displayed on the display screen is completed (performed), a "bookmark" indicating at least whether or not a review has been performed is added to each patent document that has been reviewed.

[0102] Whether or not peer review has been conducted can be determined by user actions related to the peer review process (such as input operations). For example, when sequentially switching between patent documents displayed on the screen by pressing the "Next" button, pressing the "Next" button does not indicate that peer review has been conducted. However, pressing a dedicated button (second software button) that accepts input indicating the end of peer review does indicate that peer review has been conducted. In this example, as will be described later, a "bookmark" (first bookmark) indicating the end of peer review is added to patent documents for which the peer review has been completed.

[0103] The presence or absence of a "first bookmark" assigned to each patent document allows for the identification of unreviewed patent documents, thereby preventing omissions in the review process.

[0104] Once the review of a patent document displayed on the screen is complete, the next unreviewed patent document will be displayed on the screen. For example, by operating the "Next" button, the next unreviewed patent document from the list of extracted patent documents can be displayed on the screen.

[0105] Then, from thousands of patent documents, it is determined from the description of the claims whether or not a patent document is related to the research theme, i.e., whether or not it is relevant. Based on this determination, a so-called bookmarking process is performed to record whether or not each patent document is relevant. For example, bookmarking is performed to record either no relevance or relevance for each patent document as a "bookmark" (second bookmark). After the peer review of the extracted documents (step S4) is completed, relevant patent documents are examined in the related documents review (step S5) based on the "second bookmark". For explanatory purposes, this term is called a "second bookmark," but this term can be appropriately replaced with terms such as "tag," "flag," "mark," "indicator," etc. The information held as the "second bookmark" may be binary information such as "yes or no" or "0 or 1," or it may be stepwise information indicating the degree of relevance, such as "×, ○, △, □, ◇, ☆, ●, ▲, ▼, ■, ◆, ★" or "-100, -99, ...-1, 0, 1, ..., 99, 100," and it may include textual information or numerical information. It is desirable that the information be ultimately classifiable in a binary way before the final review.

[0106] Furthermore, the "second bookmark," which indicates the presence or absence of relevance, or the degree of relevance, also functions as a "first bookmark," indicating a patent document that has undergone peer review. In other words, a patent document to which a "second bookmark" has been assigned can be considered a peer-reviewed patent document.

[0107] Here, the inventors found that, in determining relevance, irrelevant patent documents (noise documents) often contain keywords (noise words) that can serve as the basis for identifying them as noise documents, and that these noise words are sometimes common to other patent documents as well.

[0108] In determining relevance, if it is determined that there is no relevance, noise words are extracted from the patent document and stored in the computer's memory or elsewhere. When the next unreviewed patent document is displayed on the display screen of the display device, it is notified whether or not the unreviewed patent document contains noise words. If the unreviewed patent document does not contain noise words, a normal review is performed. On the other hand, if the unreviewed patent document contains noise words, it becomes easier to determine whether or not it is a noise document, i.e., whether or not it is relevance, based on the relationship between the location of the noise words and the claims. By continuing to review patent documents, noise words are accumulated. As a result, as the review of patent documents progresses, it becomes even easier to determine whether or not there is relevance. The inventor arrived at the present invention because this allows for the efficient review of patent documents.

[0109] [First Embodiment] The first embodiment will now be described. Figure 2 is a schematic diagram showing an example of the investigation support system 10 of the first embodiment. The investigation support system 10 includes a server 20 equipped with a processor connected to a network 30, and a display 21 connected to the server 20. Multiple terminal devices 40 and a patent information database (DB) 32 are connected to the network 30. The patent information database 32 is a database that stores patent information. The patent information database 32 may be either a commercial database or a Japan Patent Office database. The type of patent information database 32 is not limited here. The network 30 may be wired or wireless.

[0110] Terminal device 40 is a terminal device for using the research support system 10. Each user accesses the research support system 10 via their respective terminal device 40 and inputs the search conditions devised in the search strategy. Users can receive patent documents that match the search conditions (hits) from the server 20 as multiple documents extracted based on the set research theme, and these documents are received by the terminal device 40. Terminal device 40 may be a notebook computer, desktop computer, smartphone, or tablet, but is not limited to these.

[0111] Server 20 accepts search criteria from terminal device 40. Based on the search criteria, Server 20 searches the patent information database 32 or refers to the extracted database to extract multiple patent documents related to the research theme. The user can check the contents of the multiple patent documents extracted on terminal device 40. If the user determines that the extracted patent documents have little relevance to the research theme, the user can input new keywords and classifications via terminal device 40 and have Server 20 perform a new search. The user can formulate a search strategy while the search is being performed.

[0112] Figure 2 shows an example of the research support system 10, but is not limited to this. For example, the patent information database 32 may be provided on the server 20. Also, in the case of an in-house research support system 10, the network 30 may be configured as an internal company network.

[0113] [Server and Terminal Device Configuration] Figure 3 is a block diagram showing the schematic hardware configuration of server 20 and terminal device 40. As shown in Figure 3, server 20 and terminal device 40 have basically the same configuration. The configuration of terminal device 40 is represented by the numbers in parentheses. Server 20 is configured with a display 21, processor 22, RAM 23 (Random Access Memory), ROM 24 (Read Only Memory), storage device 25, operation unit 26, and input / output interface 27, etc. These components are connected by a bus 28. Similarly, terminal device 40 is configured with a display 41, processor 42, RAM 43 (Random Access Memory), ROM 44 (Read Only Memory), storage device 45, operation unit 46, and input / output interface 47, etc. These components are connected by a bus 48.

[0114] As previously described, the server 20 and the terminal device 40 are composed of similar hardware, but the programs executed on the processors 22 and 42 may differ, and the data stored in the storage devices 25 and 45 may also differ.

[0115] [Configuration of Input / Output Interfaces] Input / output interfaces 27 and 47 consist of terminals and slots for connecting external devices such as displays, earphones, speakers, printers, and storage media, as well as communication interfaces such as Wi-Fi (registered trademark) or Bluetooth (registered trademark). Server 20 and terminal device 40 are connected to each other via input / output interfaces 27 and 47 and network 30, enabling them to communicate with each other and perform data transmission and reception.

[0116] [Configuration of the operating unit and display] The operating units 26 and 46 are composed of devices such as a keyboard, mouse, buttons, switches, microphone, camera, motion capture device, image sensor, and electroencephalogram sensor (not shown). Users can give instructions to the server 20 via these devices. The processor 22 receives these instructions and performs processing according to the received instructions. The displays 21 and 41 can be touch panel type devices. Users may also be able to give instructions via the touch panel. The displays 21 and 41 can be composed of such touch panel type devices or liquid crystal display devices.

[0117] Display 21 displays information related to the server 20. For example, display 21 displays information related to the functions of the server 20, such as the contents of a program and the execution status of a program. Display 41 displays search results from the server 20, such as patent documents related to the research topic, and the peer review screen described later.

[0118] [Configuration of the storage device] The storage devices 25 and 45 are composed of a hard disk, a semiconductor memory such as an SSD (Solid-State Drive), a storage medium (non-temporary and tangible computer-readable medium) such as various magneto-optical storage media, and their respective control units, and store or save various types of information.

[0119] [Processor Configuration] In this embodiment, each process is executed on any computer. Furthermore, any computer may execute these processes using a processor as hardware, a program as software, or a combination thereof. In that case, the processor is configured to work in cooperation with the program to execute the various processes in this embodiment, and can function as a unit or means in this embodiment. Also, the execution order of the processes by the processor is not limited to the order described and may be changed as appropriate. Any computer may be a general-purpose computer, a computer designed for a specific purpose, a workstation, or any other system capable of executing each process.

[0120] Processors 22 and 42 may be composed of one or more hardware components, and the type of hardware is not limited. For example, a processor may be composed of hardware such as a CPU (Central Processing Unit), MPU (Micro Processing Unit), FPGA (Field Programmable Gate Array) or other programmable logic devices, an ASIC (Application Specific Integrated Circuit) or other dedicated circuitry for executing specific processes, a GPU (Graphic Processing Unit), or an NPU (Neural Processing Unit). Furthermore, the type of hardware may be a combination of different types of hardware. When multiple hardware components are configured to execute one or more processes of a given processor, these multiple hardware components may reside in physically separate devices or in the same device. Also, in any embodiment, the order of each process performed by the processor is not limited to the order described above and may be changed as appropriate. Hardware is composed of electrical circuits (circuitry) that combine circuit elements such as semiconductor elements.

[0121] Furthermore, the program may be firmware or software such as microcode. Alternatively, the program may be, for example, a set of program modules, each function of which may be implemented by a processor configured to perform its respective function. The program may also be program code or multiple code segments stored in one or more non-temporary computer-readable media (e.g., storage media or other storage; ROMs 24 and 44, or storage devices 25 and 45 (hereinafter the same)). The program may be divided and stored in multiple non-temporary computer-readable media located in physically separate devices. Program code or code segments may represent any combination of procedures, functions, subprograms, routines, subroutines, modules, software packages, classes, or instructions, data structures, or program statements. Program code or code segments may be connected to other code segments or hardware circuits by sending and receiving information, data, arguments, parameters, or memory contents.

[0122] In this embodiment, the "non-temporary and tangible computer-readable medium" does not include intangible storage media such as carrier signals or propagated signals themselves. The processor 22 can use the RAM 23 as a temporary storage area or work area when processing with a program, and the processor 42 can use the RAM 43 as a temporary storage area or work area when processing with a program.

[0123] Furthermore, the functions of processors 22 and 42 described above may be implemented by various forms of AI (Artificial Intelligence). Such AI may be, for example, an AI that performs image recognition, segmentation, or feature extraction, or an AI that generates images from given information. These AIs can also be implemented by hardware, software, firmware, microcode, or a combination thereof, as described above.

[0124] [Processor Functional Configuration Diagram] Figure 4 is a diagram showing the functional configurations of processors 22 and 42. As shown in Figure 4, processors 22 and 42 include a search condition receiving unit 220, a search processing unit 221, a noise word receiving unit 222, a noise word processing unit 223, a display screen processing unit 224, an automatic reading processing unit 225, and an input / output control unit 226.

[0125] The search condition receiving unit 220 receives the search conditions transmitted from the terminal device 40. The search conditions are entered into the terminal device 40 by the user via the operation unit 46.

[0126] The search processing unit 221 executes a search based on the search conditions. The search processing unit 221 extracts patent documents that match the search conditions from the patent information database 32. The search processing unit 221 stores the search conditions received by the search condition receiving unit 220 in association with the user's account. The search processing unit 221 also stores the search results in association with the user's account. The user can retrieve the search conditions and search results stored in the terminal device 40.

[0127] The noise word receiving unit 222 receives noise words transmitted from the terminal device 40. Noise words are input to the terminal device 40 by the user via the operation unit 46.

[0128] The noise word receiving unit 222 receives noise words based on user operations in the operation unit 46, and the noise word processing unit 223 performs processing on the noise words received by the noise word receiving unit 222. For example, the noise word processing unit 223 highlights the noise words in the patent publication. From the viewpoint of reducing processing load, it is preferable that the noise words are highlighted in the claims of the patent publication. This is because the relationship between patent documents and research themes is often based on the description of the claims in the patent documents.

[0129] Furthermore, it is preferable that the noise word processing unit 223 be able to distinguish between patent documents in which noise words are not included in the claims (first patent document) and patent documents in which noise words are included in the claims (second patent document). In this case, the noise word processing unit 223 provides identification information to at least one of the first patent document and the second patent document to distinguish between the two documents.

[0130] Multiple patent documents retrieved by the search processing unit 221 are displayed on the list screen 100 (see Figure 6), which will be described later. The identification information assigned to at least one of the first and second patent documents can be used to rearrange the first and second patent documents within the multiple patent documents. For example, the first patent document can be displayed in the front part of the list, and the first patent document can be displayed in the back part of the list.

[0131] Furthermore, the identification information assigned to at least one of the first and second patent documents can be used when selecting only the first patent document from among multiple patent documents as the patent document for peer review. For example, when displaying multiple patent documents sequentially during peer review, it can be used to skip the second patent document and select and display only the first patent document.

[0132] Furthermore, the display screen processing unit 224 can, based on the identification information assigned to at least one of the first and second patent documents, display the first and second patent documents in a distinguishable manner on the list screen 100 (see Figure 6), or, on the peer review screen 200 (see Figure 7), which will be described later, it can assign identification information to the patent documents displayed on the peer review screen 200 indicating whether the patent document is the first or second patent document.

[0133] This allows users to determine whether or not a patent document contains noise words in its claims before reviewing the document.

[0134] Furthermore, the noise word processing unit 223 performs editing such as deleting the received noise words.

[0135] The display screen processing unit 224 generates display screens to be displayed on the server 20's display 21 and the terminal device 40's display 41, and displays the display screens on displays 21 and 41. The display screens displayed on displays 21 and 41 may be the same or different. Furthermore, the display screen may be displayed on only one of displays 21 or 41. Multiple display screens may exist within the same display. The display screen on displays 21 or 41 constitutes the user interface. Through the display screen, the user can efficiently view information from the research support system 10 and perform various operations on the research support system 10. The display screens generated by the display screen processing unit 224 are customizable, improving user convenience.

[0136] The automatic reading processing unit 225 is the part that executes automatic reading software and automatically reads aloud information from the survey support system 10. This automatic reading processing unit 225 reads aloud automatically in place of or in synchronization with the information display by the display screen processing unit 224.

[0137] The information from the survey support system 10, which is automatically read aloud by the automatic reading processing unit 225, is output as audio information to earphones or speakers via the input / output interface 27 or 47, and can be heard by the user.

[0138] The user can select whether to display information from the survey support system 10 on the display screen of display 21 or 41, have it spoken through earphones or a speaker, or have the display on the display screen and the spoken audio from earphones or a speaker synchronized and executed simultaneously, by operating the operation unit 26 or 46.

[0139] The input / output control unit 226 controls the input and output of various data or signals. The display screen includes the entire window of the survey support system software, and also includes the display screen (list screen) 100 and the display screen (peer review screen) 200, which will be described later. The display screen also includes an area that can be displayed by scrolling.

[0140] [Processing in the Research Support System] Next, we will explain the flow of peer review of the extracted literature (step S4) shown in Figure 1, which is performed by the research support system 10.

[0141] Figure 5 is a flowchart showing a first embodiment of the research support method according to the present invention, and in particular shows the flowchart for the peer review of the extracted documents (step S4) in Figure 1. Here, we will explain the case where the documents are patent documents.

[0142] In Figure 5, multiple patent documents are extracted based on the research theme (step S41). By performing a database search using the research support system 10 (step S3), multiple patent documents are extracted based on the research theme.

[0143] The extracted patent documents are displayed on the display 41 of the terminal device 40. For example, the patent documents are displayed in list format, and the user can view the list of patent documents on the display 41.

[0144] In this example, the system has a list mode that displays a list of multiple patent documents on the display screen of the display 41, and a browsing mode that displays the contents of the patent documents on the display screen in a peer-readable format. The list mode and the browsing mode can be switched as needed by, for example, the user operating a soft button provided on each display screen.

[0145] Figure 6 shows an example of a display screen shown on the display 41 in list mode. In the display screen (list screen) 100 shown in Figure 6, multiple patent documents are displayed. The list screen 100 is created by the display screen processing unit 224 and displayed on the display 41. The list screen 100 displays multiple function buttons 101 and the search results performed by the search processing unit 221 in the format of list 110. List 110 includes items such as number 111, bookmark 112, application number 113, publication number 114, title of invention 115, IPC (International Patent Classification) 116, applicant 117, application date 118, publication date 119, and status 120. Number 111 is a sequential number assigned to the patent publications included in list 110. The last patent publication in list 110 is assigned the number 3000. That is, it can be understood that list 110 contains 3000 patent publications.

[0146] As described later, bookmarks 112 can be arbitrarily assigned by the user to each patent document. Bookmarks 112 can be used to check whether or not there is a relationship between the research theme and the patent document, or the degree of that relationship, after the peer review is completed. Note that the list screen 100 shown in Figure 6 is the screen before the peer review begins, and the bookmark 112 column is blank.

[0147] Application number 113 and publication number 114 function as identification numbers to specify the patent document. The filing date 118 corresponding to application number 113 and the publication date 119 corresponding to publication number 114 are included in list 110.

[0148] Publication number 114 includes the publication number and patent number (registration number). Status 120 indicates the examination and patenting status of the application relating to the patent document identified by application number 113 and publication number 114. Status 120 displays the current application status, such as application, registration, or request for examination. The title of the invention is displayed in Title of Invention 115, the International Patent Classification in IPC 116, and the applicant's name in Applicant 117.

[0149] List 110 can be sorted as needed by items included in the list, such as application number 113, publication number 114, title of invention 115, IPC 116, applicant 117, filing date 118, publication date 119, and status 120.

[0150] The multiple function buttons 101 are soft buttons provided on the list screen 100, and each can perform a different function. For example, functions such as menu screen, print, download, and change display format can be assigned to them. However, they are not limited to these. Users can customize the function buttons 101, for example, by assigning frequently used functions to the function buttons 101.

[0151] Returning to Figure 5, in step S41, list mode is set and the list screen 100 shown in Figure 6 is displayed.

[0152] Next, in step S41, when a user is instructed to change from list mode to browsing mode, the system switches from list mode to browsing mode. Specifically, when the user selects a patent document listed in list 110 on the list screen 100 using a mouse operation (left click), the system switches to browsing mode, and the selected patent document from among the multiple patent documents extracted in step S41 is displayed on the display 21 (step S42). Note that the switch from list mode to browsing mode may also be performed, for example, by operating a function button 101 set on the list screen 100, which is assigned the function of switching from list mode to browsing mode.

[0153] Unlike the list display in step S41, the display of patent documents in step S42 allows the user to view the contents of each of the extracted patent documents one by one.

[0154] Figure 7 shows an example of a display screen shown in browsing mode.

[0155] As shown in Figure 7, in browsing mode, the display 41 shows a different display screen (peer review screen) 200 than the list screen 100 shown in Figure 6. The user can review the contents of the patent document using the peer review screen 200.

[0156] The peer review screen 200 includes a function field 201, a noise word setting field 202, a bibliographic information field 203, an abstract field 204, a claims field 205, and a drawing field 206, among others. The noise word setting field 202 is an example of an input reception field for the present invention.

[0157] The function column 201 contains information and functions related to patent documents. The "Number:" column displays a number corresponding to number 111 in list 110. By looking at the number, the user can find out which patent document is being displayed in list 110.

[0158] The "List" button 201A is a function button for returning to the list screen 100. The "Previous" button 201B is a function button for displaying the previous patent document for the currently displayed patent document. The "Next" button 201C is a function button for displaying the next patent document for the currently displayed patent document. The "Print" button 201D is a function button for printing the currently displayed patent document. When the "Print" button 201D is pressed, the user can be taken to a screen where they can specify the printing style, such as printing the publication or printing the front page.

[0159] Furthermore, the function section 201 is provided with a "bookmark" button (in this example, a soft button consisting of three shapes (□, ○, △) (third soft button)) 201E for adding a "bookmark" (second bookmark) to the patent document, indicating the result of determining the relationship between the patent document and the research theme displayed on the peer review screen 200.

[0160] By selecting one of the three "bookmark" buttons 201E (□, ○, △), users can add a "bookmark" (second bookmark) to the patent document displayed on the peer review screen 200, indicating the result of determining the relationship between the research theme and the patent document.

[0161] For example, if the three "bookmark" buttons 201E (□, ○, △) are each assigned a "second bookmark" indicating three levels of relevance: low, medium, and high, then when the "bookmark" button 201E for □ is operated, the patent document displayed on the peer review screen 200 will be given a "second bookmark" indicating low relevance as a result of relevance determination. When the "bookmark" button 201E for △ is operated, the patent document will be given a "second bookmark" indicating high relevance as a result of relevance determination. When the "bookmark" button 201E for ○ is operated, the patent document will be given a "second bookmark" indicating an intermediate level of relevance between high and low as a result of relevance determination.

[0162] Furthermore, the "bookmark" button 201E (third soft button) used to indicate the determination of the relationship between the patent document and the research theme is not limited to the three mentioned above; it may be four or more buttons indicating different degrees of relationship, or it may be two buttons indicating whether or not there is a relationship.

[0163] Furthermore, when the cursor moves over the area containing text ("List", "Previous", "Next", "Print") or shapes (□, ○, △) on the function area 201, the "List" button 201A, "Previous", "Next", "Print" button 201B, "Next", "Print

[0164] The functions of function section 201 are not limited to these. Users can customize function section 201, for example, by adding function buttons or assigning other functions to existing function buttons.

[0165] The noise word setting field 202 is a field where the user enters terms that they have determined to be noise words. By using the noise word setting field 202, the user can easily enter noise words. By displaying the noise word setting field 202 on the peer review screen 200, the user can enter noise words at any time during the peer review process. It is desirable that there be a prompt to enter noise words within or outside the noise word setting field 202, for example, a description such as "Noise word setting field," "Enter noise words," "Tips: Highlighting noise words makes the publication easier to read," or "Type noise words." It is also acceptable to display symbols or shapes such as "!" on the screen, and when the mouse cursor is placed over the symbol or shape, a prompt to enter the above-mentioned noise words may pop up.

[0166] Furthermore, the noise word setting field 202 in this example is provided with a "Noise Word Setting" button 202A. The "Noise Word Setting" button 202A is a registration software button (first software button) that stores the noise words entered in the noise word setting field in the storage device 25 in association with the multiple patent documents extracted in step S41. After entering an unregistered noise word in the noise word setting field 202, the user can add the unregistered noise word to the storage device 25 by operating the "Noise Word Setting" button 202A.

[0167] Alternatively, the "Noise Word Setting" button 202A may be replaced with a "Highlight" button. This is because noise words registered in the storage device 25 are highlighted (including highlighted) in the patent documents displayed on the peer review screen 200.

[0168] The bibliographic information section 203 displays information and additional information corresponding to the items in list 110. For example, it displays the application number, publication number, filing date, publication date, title of the invention, IPC, FI, applicant, and inventor.

[0169] The abstract section 204 displays the abstract described in the patent document. The claims section 205 displays the claims described in the patent document, and the drawings section 206 displays the drawings described in the patent document. Although not shown in the figure, the user can view the specification, which includes a detailed description of the invention in the displayed patent document, by scrolling the display screen 200.

[0170] Next, the user determines whether or not there are any unregistered (new) noise words in the patent document (step S43 in Figure 5). That is, the user extracts unregistered noise words from the patent publication that will serve as the basis for determining that the patent document and the research theme are unrelated. If a noise word is highlighted, it is a registered noise word that has already been registered in the storage device 25.

[0171] Furthermore, if the patent publication contains noise words, the function section 201 may display "Noise words present". Even if the noise words are not present within the display field of view, they may be hidden within the patent publication. Users who see "Noise words present" in the function section 201 can find the noise words by scrolling the review screen 200. The display of "Noise words present" serves as a warning to the user. The display of "Noise words present" is made by the noise word processing unit 223 to display it on the display screen 200 via the display screen processing unit 224.

[0172] If there are unregistered noise words that are not highlighted (if the answer is Yes in step S43), the user enters the unregistered noise words into the noise word setting field 202 and operates the "Noise Word Setting" button 202A to store the unregistered noise words in the storage device 25 (step S44).

[0173] Figure 8 shows an example of a display screen shown in browsing mode, and specifically illustrates the case where unregistered noise words are stored in the storage device 25.

[0174] As shown in Figure 8, when a user extracts an unregistered noise word, they input the extracted noise word into the noise word setting field 202. In this case, the user can move the cursor to a predetermined position in the noise word setting field 202 and input the word by operating the keyboard or other means. In the example shown in Figure 8, the unregistered noise word to be registered (highlighted) is "red blood cell".

[0175] Next, the user operates the "Noise Word Setting" button 202A to register the noise words entered in the noise word setting field 202 into the storage device 25. The "Noise Word Setting" button 202A can be operated by moving the cursor (pointer) over the "Noise Word Setting" button 202A and clicking the mouse button, or by pressing the Enter key on the keyboard.

[0176] Once noise words are registered in the storage device 25 in this manner, the noise word processing unit 223 highlights the registered noise words in the patent publication. This allows users to easily see the registered noise words included in the patent documents displayed on the peer review screen 200, and also to search for unregistered noise words.

[0177] Methods for highlighting noise words include highlighting, coloring, bolding, and underlining. This allows users to easily recognize noise words. In Figure 8, noise words are highlighted with a rectangular frame. If multiple noise words are registered, each noise word may be highlighted in a different way (for example, a different color in the case of colored text).

[0178] Noise words can be entered into the noise word setting field 202 by user operations on the operation unit 46, such as keyboard input, voice input, copy and paste, or drag and drop. User operations on the operation unit 46 also include selecting and registering unregistered noise words using the review screen 200, which allows users to review the contents of patent documents, and these operations make it easy to register unregistered noise words.

[0179] Here, the selection of noise words using the peer review screen includes cursor operations to select unregistered noise words on the peer review screen. Cursor operations can be performed using mouse operations, eye tracking operations, gesture operations, touchpad operations, etc.

[0180] The operation to register a selected, unregistered noise word is an instruction operation that causes the noise word selected by cursor operation to be stored in the storage device 25 (to input to the noise word receiving unit 222). This instruction operation can be performed using mouse clicks (left click, right click, wheel click, double click), touchpad taps, keyboard, voice, eye tracking, gestures, and electroencephalograms. However, if the basis for determining that a document is a noise document is something other than a noise word, for example, if the claims description contains an extremely large number of constituent elements and appears unrelated at first glance, there are cases where a noise word will not be extracted.

[0181] Furthermore, when entering an unregistered noise word in the noise word setting field 202, it is necessary to move the cursor (pointer) to a predetermined position in the noise word setting field 202. Here, the predetermined position is the beginning of the noise word setting field 202 if there are no registered noise words and the noise word setting field 202 is blank, and if there are registered noise words and registered noise words are displayed in the noise word setting field 202, it is the position one space after the last registered noise word displayed in the noise word setting field 202.

[0182] When moving the cursor to the designated position in the noise word setting field 202, this can be done using cursor operations with a mouse or the like, but it may also be done by using a shortcut key (first shortcut key) assigned to the keyboard.

[0183] Specifically, a first shortcut key is assigned to any key on the keyboard constituting the operation unit 46. This first shortcut key enables the input of noise words into the noise word setting field 202, and when the first shortcut key is pressed, it moves the cursor on the review screen 200 to a predetermined position in the noise word setting field 202. As a result, the cursor can be moved to a predetermined position in the noise word setting field 202 by simply pressing the first shortcut key assigned to the keyboard once, thereby shortening the time required to input noise words.

[0184] Furthermore, in this example, after entering a noise word in the noise word setting field 202, operating the "Noise Word Setting" button 202A stores the entered noise word in the storage device 25, and also highlights the noise word included in the patent document displayed on the peer review screen 200. This makes it easy to distinguish between the noise word and other words on the peer review screen 200.

[0185] To operate the "Noise Language Settings" button 202A, you need to move the cursor over the "Noise Language Settings" button 202A. You can move the cursor over the "Noise Language Settings" button 202A using a mouse or other cursor, but you can also do so by using a shortcut key (second shortcut key) assigned to the keyboard.

[0186] In other words, a second shortcut key is assigned to any key on the keyboard that constitutes the operation unit 46. This second shortcut key moves the cursor to the "Noise Word Setting" button 202A, and when the second shortcut key is pressed, it moves the cursor on the review screen 200 (Noise Word Setting field 202) to the "Noise Word Setting" button 202A. As a result, the cursor can be moved to the "Noise Word Setting" button 202A with just one press of the second shortcut key assigned to the keyboard, and the operation of the "Noise Word Setting" button 202A can be performed immediately.

[0187] Returning to Figure 5, if there are no unregistered noise words (the result is No in step S43), or if an unregistered noise word is registered in the storage device 25 in step S44, the process proceeds to step S45, where it is determined whether or not the review of the patent document displayed on the review screen 200 has been completed.

[0188] If the review of the patent document displayed on the review screen 200 is not complete (indicated as No in step S44), the process returns to step S42.

[0189] In step S42, if an unregistered noise word is added to the storage device 25 in step S44, that unregistered noise word will be highlighted. This allows the user to confirm that an unregistered noise word has been added to the storage device 25. Furthermore, by repeating the process from step S43 to step S45, it is possible to register two or more different unregistered noise words based on the same patent document.

[0190] On the other hand, if the review of a patent document displayed on the review screen 200 is completed (if the answer is Yes in step S44), the user adds a "bookmark" ("first bookmark" or "second bookmark") to the patent document that has been reviewed (step S46).

[0191] The "first bookmark" is displayed on the peer review screen 200 and is assigned to each patent document that has been reviewed. Patent documents to which the "first bookmark" has been assigned can be identified as reviewed patent documents.

[0192] Furthermore, the "second bookmark" is a bookmark that indicates whether or not there is a relationship between the patent document displayed on the peer review screen 200 and the research theme, or the degree of that relationship (the result of the user's determination of the relationship), and is assigned to each patent document displayed on the peer review screen 200.

[0193] Figure 9 shows an example of a display screen shown in browsing mode, and in particular shows the state in which the "bookmark" button 201E has been operated and a "second bookmark" has been added to the patent document displayed on the peer review screen 200.

[0194] In step S46 of Figure 5, the user determines the relationship between the patent document displayed on the peer review screen 200 and the research theme, and selects one of the three shapes (□, ○, △) of the "bookmark" button 201E. Based on the user's operation of the "bookmark" button 201E, the processors 22 and 42 receive a determination result indicating whether or not there is a relationship, or the degree of the relationship, and add a "second bookmark" indicating the received relationship determination result to the patent document displayed on the peer review screen 200.

[0195] In the example shown in Figure 9, a "second bookmark" indicating low relevance is added to the patent document. The "bookmark" button 201E (square) is operated, and the "second bookmark" indicating low relevance between the patent document and the research topic is added to the patent document. The color of the "bookmark" button 201E (square, circle, triangle) is changed to a square. Users can check the relevance between the patent document and the research topic based on which of the three "bookmark" buttons 201E (square, circle, triangle) has had its color changed.

[0196] Furthermore, when a "second bookmark" is assigned to each patent document, as shown in Figure 11, the information of the "second bookmark" assigned to each patent document (in this example, one of the shapes ■, ●, or ▲) is added to the bookmark 112 column on the list screen 100.

[0197] As shown in Figure 9, bookmarking is not limited to when the user operates the "bookmark" button 201E. Alternatively, two or more shortcut keys for "bookmarking" can be set on the keyboard, allowing bookmarking to be performed with a single press of a shortcut key, or bookmarking can be performed using voice input or electroencephalography (EEG) input.

[0198] After reviewing all extracted patent documents, users can use the "second bookmark" to select patent documents that should be examined further.

[0199] Furthermore, since the "second bookmark" is assigned to each patent document, it also functions as the "first bookmark" (indicating whether or not it has been peer-reviewed).

[0200] Next, it is determined whether or not the peer review of the multiple patent documents (all patent documents) extracted in step S41 of Figure 5 has been completed (step S47).

[0201] If the peer review of all patent documents has not been completed (if the flag indicating completion of peer review has not been set) (if the result in step S47 is No), the system returns to step S42 and the user operates the "Next" button 201C to display the next patent document number.

[0202] The conditions for setting the flag indicating the end of the review process are that all extracted documents have been reviewed, a user-defined number of documents have been reviewed, or the reviewer themselves can manually set the flag. Setting the flag allows reviewers to confirm that the review process is complete, thus preventing accidental re-reviews. Furthermore, it allows detailed reviewers to conduct thorough examinations with the assurance that the review process is truly complete.

[0203] The user repeatedly performs the determination and processing between steps S42 and S47 until the review of all patent documents is completed (until the result is "Yes" in step S47).

[0204] Figure 10 shows a state where peer review of multiple patent documents has been completed to a certain extent. In this example, the user can add noise words as needed. In Figure 10, multiple noise words are highlighted.

[0205] Figure 12 is a diagram showing a portion of a flowchart illustrating the first embodiment, and shows step S46-1, which replaces step S46 shown in Figure 5.

[0206] In Figure 12, the user determines whether or not there are noise words in the patent documents displayed on the peer review screen 200 that have been reviewed (step S50).

[0207] If there are no noise words (in the case of No in step S50), the relationship between the patent document and the research theme is determined by comparing the content of the patent document with the research theme (step S51). For example, the user reads the claims (especially claim 1) described in the claims section 205 of the peer review screen 200 to check the content and determine whether or not there is a relationship with the research theme. Since the patent document does not contain noise words, the user carefully compares the claims with the research theme to make a determination.

[0208] Furthermore, patent documents that do not contain noise words are more likely to be subject to scrutiny (and conversely, less likely to be considered noise documents), and the peer review process required to determine the relationship between the two takes a relatively long time.

[0209] On the other hand, if there are noise words (if the answer is Yes in step S50), the user refers to the noise words to determine the relationship between the patent document and the research topic (step S52). In some cases, the presence of even one noise word can be used to determine that there is no relationship, significantly reducing the review time. Furthermore, patent documents containing many noise words have a higher probability of being noise documents, allowing for quick determination that the patent document and the research topic are unrelated. As the review of the patent document progresses, the number of registered noise words increases, so as the review of the patent document nears completion, it becomes possible to determine that the patent document and the research topic are unrelated in an even shorter amount of time.

[0210] In other words, in this example, by highlighting noise words, it is possible to visualize the parts that could be the basis for noise in the literature, thereby minimizing the time spent searching for patent documents while maintaining accuracy.

[0211] If the relationship between the patent document and the research theme is determined in step S51 or step S52, a bookmark indicating the determination result is placed on the patent document (a "second bookmark" is added) (step S53).

[0212] Figure 11 shows the list screen 100 after the review of patent documents has been completed. As shown in Figure 11 and in List 110, the items in Bookmark 112 are marked with a graphic symbol (■, ●, or ▲) indicating the result of bookmarking. Patent documents marked with a ▲ are highly relevant to the research theme and are therefore subject to detailed examination in step S5 of Figure 1. The research support system 10 may have a function to output the list of patent documents marked with a ▲ as data, as shown in Figure 11. It may also have a function to output a list of extracted noise words.

[0213] In this example, the noise word setting field 202 is shown on the review screen 200, but this is not limited to this. For example, instead of providing a noise word setting field 202, a noise word setting button may be displayed in the function field 201. When the user operates the noise word setting button, a noise word input window may open and be displayed on the review screen 200.

[0214] While the example illustrates the addition of noise words, users can also edit them. For example, users can directly add, modify, or delete noise words on the noise word setting field 202, or they can delete or modify noise words. For example, users can edit noise words by switching the functionality of the noise word setting field 202 on the peer review screen 200 to add mode, delete mode, and modify mode. Accepting editing of noise words makes it possible to add, modify, or delete noise words, and as a result, the relationship investigation between literature and research themes becomes more accurate.

[0215] Noise words can also be entered via voice input into the noise word setting field 202. Users can input noise words by voice. In this case, the user can first select the editing mode by saying one of the following: "Add Mode," "Delete Mode," or "Modify Mode."

[0216] Regarding the input of noise words into the noise word setting field 202, any method to improve efficiency can be applied as needed. For example, when a user inputs a noise word, it may be highlighted sequentially, i.e., in real time, without being confirmed by pressing the Enter key.

[0217] The peer review screen 200 illustrates the case where one patent document is displayed. In this example, the displayed patent document is changed by switching the peer review screen 200 using the "Previous" button 201B and the "Next" button 201C in the function field 201. When switching between the "Previous" and "Next" displays of the patent document, the user may press the "Previous" button 201B or the "Next" button 201C, or they may use keyboard input, voice input, eye-tracking input, gesture input, electroencephalogram (EEG) input, etc.

[0218] The peer review screen 200 is not limited to the above and can also display multiple patent documents in sequence. Multiple patent documents can be checked for the presence or absence of noise words in sequence. In this example, the displayed patent documents are changed by scrolling the peer review screen 200. The number of patent documents to display in sequence can also be set. For example, if the number is set to 50, the initial display screen 200 allows viewing of patent documents numbered 1 to 50 by scrolling. After viewing patent document number 50, the "Next" function button allows viewing of patent documents numbered 51 to 100 by scrolling on the next peer review screen 200. It is important that multiple document data can be displayed immediately; therefore, instead of scrolling, a method that caches multiple document data to suppress display delays while switching the peer review screen 200 for each document is also acceptable. It is desirable that the peer review screen 200 does not reload when highlighting is performed.

[0219] In the research support system 10 of this example, the peer review screen 200 for viewing patent documents displays a noise word setting field 202. The research support system 10 can select whether or not to perform peer review using the noise word setting field 202.

[0220] For example, when performing a database search as shown in Figure 1 (step S3), if the research support system 10 is used, it is not necessary to use the noise word setting field 202. Therefore, a menu named "Noise Literature Review" is provided on the menu screen of the research support system 10. Users can decide whether or not to use "Noise Literature Review" to register noise words.

[0221] Furthermore, it is preferable that the user manual include a statement such as, "When using the 'Noise Removal and Review' function, please enter the noise word that caused you to determine the document to be noisy during the review process in the noise word setting field. If that noise word appears in any unreviewed documents, it will be highlighted..."

[0222] [Second Embodiment] Figure 13 is a diagram showing a part of a flowchart for explaining the second embodiment, and shows step S46-2 which replaces step S46 shown in Figure 7.

[0223] In Figure 13, the parts that are common to the flowchart shown in Figure 12 are given the same step numbers.

[0224] In step S50 of Figure 13, if the user determines that there are no noise words in the patent documents displayed on the peer review screen 200 that have been reviewed (in the case of No in step S50), the relationship between the patent document and the research theme is determined by comparing the content of the patent document with the research theme (step S51), and a bookmark indicating the determination result is added to the patent document (a "second bookmark" is added) (step S53).

[0225] On the other hand, in step S50 of Figure 13, if the user determines that there is a noise word in the patent document (if the answer is Yes in step S50), the user then determines whether there are any expressions requiring attention near the noise word contained in the patent document (for example, alternative statements, conditional branching sections, and negative expressions) (step S60).

[0226] Generally, if even one noise word is present in a patent document, it is highly likely to be identified as a noise document. On the other hand, even if a noise word is present in a patent document, it may not be identified as a noise document depending on the terminology written near the noise word. For example, this may occur if there are other expressions of concern near the noise word that are different from the noise word and affect the determination of whether or not the patent document is related to the research topic.

[0227] Let's assume the noise term is "resin" and the term for the research topic is "metal." A patent claim in a patent document might state, "The first component is resin or metal." Because the patent document includes "resin," it might appear at first glance to be noise literature. However, due to its alternative wording, this patent document also includes the term "metal" from the research topic. Distinguishing a document as noise literature based solely on the noise term carries the risk of overlooking important information.

[0228] Furthermore, if the claims of a patent document include a "conditional branch" where the actions performed differ depending on the condition, such as "if A, then B; if not A, then C," then if the research topic satisfies any of the "conditional branch," it can be determined that it is not noise literature. If it does not satisfy any of the "conditional branch" (for example, "if A, then C; if not A, then B"), it can also be determined that it is not noise literature.

[0229] Furthermore, patent claims in patent documents may contain "negative expressions." Here, "negative expressions" refer to expressions that are understood to indicate negation, exclusion, absence, non-existence, etc., and include words and affixes such as "not," "non-," "other than," "excluding," "none," and "not," or specific syntactic patterns.

[0230] Let's assume the noise term is "resin" and the term for the research topic is "metal." If the claims of a patent document include descriptions of "non-metallic" or "materials other than metals," these terms do not refer to "resin," but they negate or exclude "metals" from the research topic, thus allowing the document to be identified as noise.

[0231] Therefore, in the second embodiment, in step S60, it is determined whether or not there is an expression requiring attention (for example, an alternative statement, a conditional branch, or a negative expression) near the noise word.

[0232] Here, the vicinity of a noise word can be arbitrarily defined by the user as, for example, the five words before and after the noise word, that is, from the five words before the comma preceding the noise word to the five words after the comma following the noise word.

[0233] Examples of alternative descriptions include "or," "or," "or," "and," "or," "alternatively," "either...or...," "and / or," " / ," and "as an alternative." Furthermore, examples of alternative descriptions include marker-style descriptions, such as "selected from the group consisting of..." and "selected from the group consisting...."

[0234] If there are no expressions of concern (alternative statements, conditional branching sections, or negative expressions) near the noise word (i.e., if the result is No in step S60), proceed to step S45. Since the patent document does not contain any expressions of concern, it is highly likely to be a noise document. Therefore, it is preferable to remember that there are no expressions of concern. The user can use this as a reference when bookmarking in step S53. As a method of remembering, a flag can be stored in memory, or bookmarking can be done prior to bookmarking in step S53.

[0235] If there are no expressions of concern near the noise word (in the case of No in step S60), the noise word is referenced to determine the relationship between the patent document and the research theme (step S52). If the relationship between the patent document and the research theme is determined in step S52, the process proceeds to step S53, where a bookmark indicating the determination result is added.

[0236] On the other hand, if there is a cautionary expression near a noise word (in the case of Yes in step S60), the distinguishability between the noise word near the cautionary expression and other words or phrases is reduced (step S62).

[0237] In other words, normally, noise words are highlighted on the peer review screen 200, but if there is a noteworthy expression near the noise word, the visibility of the noise word due to the highlighting is reduced, making the noise word less conspicuous. The reduction in distinguishability between the noise word and other words includes making them indistinguishable (not highlighting the noise word).

[0238] The reason for reducing the identifiability of noise words in this way is that when noise words are highlighted, users are more likely to use their presence as evidence to determine that the patent document under review is unrelated to the research topic.

[0239] Furthermore, in step S62, it is preferable to reduce the identifiability of noise words and make it possible to distinguish between expressions requiring attention and other words. In this case, it is preferable to display "noise words" and "expressions requiring attention" in an identifiable manner. This is to make it easy to distinguish between "noise words" and "expressions requiring attention."

[0240] Next, the user refers to noise words and expressions of concern to determine the relationship between the patent document and the research topic (step S63). Patent documents that contain noise words in their claims (especially claim 1) are usually more likely to be identified as noise documents. However, if there are expressions of concern, such as alternative descriptions, near the noise words, it may not be desirable to treat them as noise words.

[0241] Therefore, if there is a noteworthy expression near a noise word, it is preferable to refer to the noteworthy expression together with the noise word in order to more accurately determine the relationship between the patent document and the research topic.

[0242] If the relationship between the patent document and the research topic is determined in step S63, the process proceeds to step S53, where a bookmark indicating the determination result is created.

[0243] In the second embodiment, step S60 can be performed by the noise word processing unit 223. The user registers the definition of "neighborhood" and "expressions requiring attention" ("alternative description", "conditional branching section", or "negative expression") as conditions to the noise word processing unit 223. Based on the conditions, the noise word processing unit 223 determines whether or not there are expressions requiring attention in the vicinity of the noise word. If the noise word processing unit 223 determines that there are no expressions requiring attention in the vicinity of the noise word, it stores a flag in memory. By having the noise word processing unit 223 perform the determination in step S51, the time required to identify noisy documents can be significantly reduced.

[0244] Furthermore, the terms requiring attention are not limited to those located near noise words; for example, if a term related to the research theme, or a term similar to the research theme, is extracted from a patent document, the terms requiring attention may also include those located near the extracted term.

[0245] For example, if a patent document does not contain any noise words and does contain terms related to the research topic, that patent document can usually be considered a document relevant to the research topic (a document that should be examined closely). However, if there is a cautionary expression near the terms related to the research topic, and that cautionary expression negates the terms related to the research topic, then it may be possible to consider it a document that is not relevant to the research topic.

[0246] In the second embodiment, the method can be applied not only when the user registers noise words as in the first embodiment, but also when noise words are registered in advance before peer review.

[0247] Users register noise words before peer review. If a noise word is found in a patent document, the system checks for any expressions of concern (alternative statements, conditional branching, or negative expressions) near the noise word. By considering the noise words and expressions of concern, it is possible to determine whether the patent document is related to the research topic.

[0248] [Automatic Reading Processing Unit] In the embodiments of the research support method shown in Figures 5, 12, and 13, the user reviews the patent documents displayed sequentially on the review screen 200 shown in Figure 7, etc., bookmarks each patent document, and registers noise words using the noise word setting field 202 on the review screen 200. However, the user is not limited to this, and may listen to the patent documents (especially the claims) automatically read aloud by the automatic reading processing unit 225, bookmark each patent document, and register noise words, for example, by voice input, without using the noise word setting field 202.

[0249] The user can use the microphone, which functions as the control unit 46, to give various instructions, including the automatic reading of patent documents, via voice input.

[0250] Upon receiving a command to automatically read aloud patent documents, the automatic reading processing unit 225 executes the automatic reading software and sequentially reads aloud multiple patent documents found by the search processing unit 221, outputting the audio information to earphones or speakers via input output 025027 or 47. The user can then review the patent documents by listening to the audio emitted from the earphones or speakers.

[0251] When the automatic reading processing unit 225 automatically reads a patent document, if the patent document contains the same noise word as a noise word registered in the storage device 25, it switches the automatic reading method for that noise word and other words.

[0252] For example, by making the timbre of "noise words" different from that of other words, making the volume of "noise words" louder than that of other words, or by slowing down the reading speed of "noise words" compared to that of other words, it becomes possible to distinguish between "noise words" and other words.

[0253] During the automatic reading of patent documents, the user can determine the relevance between the patent document being read and the research topic. For example, if it is determined that there is no relevance, the user can terminate the reading of the current patent document and instruct the user to start reading the next patent document via voice input or other means.

[0254] Furthermore, if an unregistered noise word is heard during the automatic reading of a patent document, the user can interrupt the automatic reading of the patent document and register the unregistered noise word in the storage device 25 via voice input or other means. In this case, the noise word can be prevented from being incorrectly registered by an answer-back function that repeats the voice input of the noise word.

[0255] [Effects] The effects of the embodiment are described below. The investigation time before and after highlighting noise words was compared. The comparison of investigation time was based on the estimated total time for review and scrutiny.

[0256] Before highlighting noisy terms, 100 patent documents were reviewed in one hour. The review speed was 100 documents per hour. As a result of the review, 50 patent documents were deemed suitable for further examination. In other words, the noisy document extraction rate was 50% (noisy documents / total documents = (100 - 50) documents / 100 documents).

[0257] After highlighting the noisy words, a peer review was conducted on 342 patent documents in 3 hours. The review speed was 114 documents per hour. As a result of the review, 14 patent documents were deemed suitable for further examination. In other words, the noisy document extraction rate was 96% (noisy documents / total documents = (342 - 14) documents / 342 documents).

[0258] Next, we examined 1,000 patent documents, using a review speed of 5 minutes per document, and calculated the estimated time for each.

[0259] Before highlighting noise terms, the review time is (1 hour / 100 documents) x 1000 documents = 10 hours. Furthermore, since the noise literature extraction rate is 50%, 500 documents will be subject to further examination. The examination time will be 500 documents x 5 minutes / document = approximately 42 hours. Therefore, the total estimated time is 10 hours + 42 hours = 52 hours. Assuming an 8-hour workday, the estimated research time before highlighting noise terms is 6.5 days / 1000 documents.

[0260] After highlighting noise terms, the review time will be (1 hour / 114 items) x 1000 items = 8.8 hours. Furthermore, since the noise literature extraction rate is 96%, 40 items will be subject to further examination. The examination time will be 40 items x 5 minutes / item = approximately 3.3 hours. Therefore, the total estimated time will be 8.8 hours + 3.3 hours = approximately 12 hours. The estimated research time after highlighting noise terms is 1.5 days / 1000 items.

[0261] As previously mentioned, comparing the survey time, it is 6.5 days / 1000 cases versus 1.5 days / 1000 cases, indicating a reduction of more than four times the time (a reduction of 5 days per 1000 cases). In addition, since the extracted noise words can be recorded, it will be possible to review and share the criteria for extracting noise literature after the survey.

[0262] [Note] The above embodiments are disclosed and include the following investigation support methods and investigation support systems.

[0263] (1) A research support method that extracts multiple documents based on a set research theme, allows the user to sequentially review the multiple documents, and provides each document with at least one of two bookmarks: a first bookmark indicating whether or not a review has been conducted, and a second bookmark indicating the result of determining the relationship between the document and the research theme.

[0264] (2) The research support method described in (1), wherein when a user sequentially reviews multiple documents, the multiple documents are displayed sequentially on the display screen, and noise words stored in the memory device are highlighted on the display screen for the multiple documents displayed sequentially on the display screen.

[0265] (3) Multiple documents are multiple patent documents, the search support method described in (1) or (2).

[0266] (4) The search support method described in (2), wherein identification information is provided to at least one of the first patent document and the second patent document to distinguish between a first patent document in which the noise word is not included in the claims and a second patent document in which the noise word is included in the claims, and the identification information is information that can be used when rearranging the first patent document and the second patent document in a plurality of patent documents, or when selecting only the first patent document from a plurality of patent documents as a patent document for peer review.

[0267] (5) The investigation support method described in (2), which accepts noise words to be stored in a memory device from an input reception field.

[0268] (6) A research support system equipped with a processor, wherein the processor extracts multiple documents based on a set research theme, and when the user sequentially reviews the multiple documents, it receives whether or not a review has been conducted on the documents, or receives a determination result indicating whether or not there is a relationship between the documents and the research theme, or the degree of the relationship, and assigns a first bookmark to each document indicating whether or not a review has been conducted, and a second bookmark indicating the determination result of the relationship.

[0269] (7) The research support system described in (6), wherein the processor displays multiple documents sequentially on a display screen when a user sequentially reviews multiple documents, and highlights noise words stored in the memory device on the display screen for the multiple documents displayed sequentially on the display screen.

[0270] (8) The research support system according to (7), wherein the processor assigns identification information to at least one of the first patent document and the second patent document that distinguishes between a first patent document in which the noise word is not included in the claims and a second patent document in which the noise word is included in the claims, the identification information being information that can be used when rearranging the first patent document and the second patent document in a plurality of patent documents, or when selecting only the first patent document from a plurality of patent documents as a patent document for peer review.

[0271] (9) The investigation support system according to (7) or (8), wherein the processor displays an input field for inputting noise words on a display screen and accepts noise words to be stored in a memory device from the input field.

[0272] 10 Research support system 20 Server 21 Display 22 Processor 23 RAM 24 ROM 25 Storage device 26 Operation unit 27 Input / output interface 28 Bus 30 Network 32 Patent information database 40 Terminal device 41 Display 42 Processor 43 RAM 44 ROM 45 Storage device 46 Operation unit 47 Input / output interface 48 Bus 100 Display screen (list screen) 101 Function buttons 110 List 111 Number 112 Bookmark 113 Application number 114 Publication number 115 Title 116 IPC 117 Applicant 118 Application date 119 Publication date 120 Status 200 Display screen (peer review screen) 201 Function field 201A "List" button 201B "Previous" button 201C "Next" button 201D "Print" button 201E "Bookmark" button 202 Noise word setting field 202A "Noise word setting" button 203 Bibliographic information field 204 Abstract field 205 Claim field 206 Drawing field 220 Search condition reception unit 221 Search processing unit 222 Noise word reception unit 223 Noise word processing unit 224 Display screen processing unit 225 Automatic reading processing unit 226 Input / output control unit

Claims

1. A research support method comprising: extracting multiple documents based on a set research theme; having the user sequentially review the multiple documents; determining whether the documents contain unregistered noise words; storing the unregistered noise words in the computer's memory device when input of the unregistered noise words is received; and enabling the identification of the noise words stored in the memory device from other words when multiple documents are reviewed.

2. The research support method according to claim 1, wherein, when a user sequentially reviews multiple documents, the multiple documents are displayed sequentially on a display screen, and the noise words stored in the storage device are highlighted on the display screen in the multiple documents displayed sequentially on the display screen.

3. The research support method according to claim 1, wherein when a user sequentially reviews multiple documents, the multiple documents are automatically read aloud in order, and the method of automatically reading aloud the noise words stored in the storage device and words other than the noise words in the multiple documents that are automatically read aloud in order.

4. The search support method according to claim 1, wherein the aforementioned multiple documents are multiple patent documents.

5. The search support method according to claim 4, wherein, in order to make the noise word and phrases other than the noise word distinguishable, the noise word and phrases other than the noise word are distinguishable at least within the claims of the patent document.

6. The search support method according to claim 4, which enables the identification of a first patent document in which the noise word is not included in the claims and a second patent document in which the noise word is included in the claims.

7. The research support method according to claim 4, wherein identification information is provided to at least one of the first patent document and the second patent document to distinguish between a first patent document in which the noise word is not included in the claims and a second patent document in which the noise word is included in the claims, and the identification information is information that can be used when rearranging the first patent document and the second patent document in a plurality of patent documents, or when selecting only the first patent document from a plurality of patent documents as a patent document for peer review.

8. The investigation support method according to claim 1, wherein the noise word to be stored is received from an input reception field.

9. The research support method according to claim 8, wherein the input reception field is displayed on a display screen that sequentially displays a plurality of the aforementioned documents.

10. The investigation support method according to claim 8, which accepts editing of the noise words.

11. The investigation support method according to claim 9, wherein the display screen displays a message prompting the input of noise words.

12. The investigation support method according to claim 1, wherein the noise words to be stored are received based on user operations on the operation unit.

13. The research support method according to claim 12, wherein the user operation in the operation unit is an input operation and a registration operation of an unregistered noise word performed using a display screen that sequentially displays a plurality of the documents.

14. The investigation support method according to claim 13, wherein the input operation includes a cursor operation to select the noise word on the display screen, and the registration operation is an instruction operation to store the noise word selected by the cursor operation in the storage device.

15. The investigation support method according to claim 14, wherein the cursor operation includes at least one of mouse operation, eye tracking operation, gesture operation, and touchpad operation, and the instruction operation includes at least one of mouse click operation, touchpad tap operation, keyboard input, voice input, eye tracking input, gesture input, and electroencephalogram input.

16. The investigation support method according to claim 13, wherein the display screen is provided with an input field into which the noise word to be memorized is entered, and the input operation is the operation of entering the noise word into the input field using a keyboard, voice input, or copying and pasting or dragging and dropping into the input field.

17. The investigation support method according to claim 1, wherein the noise words to be stored are received based on voice input.

18. A research support method according to any one of claims 1 to 17, which determines whether or not a noteworthy expression different from the noise word exists in the document, and, if such a noteworthy expression exists, makes it possible to distinguish the noteworthy expression from other words or phrases.

19. A research support method according to any one of claims 1 to 17, which determines whether a noteworthy expression different from the noise word exists in the document, and if the noteworthy expression exists in the vicinity of the noise word, reduces the distinguishability between the noise word and other words or phrases.

20. The investigation support method according to claim 18, wherein the cautionary expression includes at least one of an alternative statement, a conditional statement, a conditional branch, and a negative expression.

21. A research support method according to any one of claims 1 to 17, wherein when a user sequentially reviews multiple documents, the method receives information on whether or not a review has been conducted for each document, and provides each document with at least one first bookmark indicating whether or not the review has been conducted.

22. The research support method according to claim 21, wherein when a user sequentially reviews multiple of the aforementioned documents, the method determines the relationship between the documents and the research theme, and provides each of the documents with a second bookmark indicating the result of the relationship determination.

23. The research support method according to claim 22, wherein if the document does not contain the noise word, the relationship between the document and the research theme is determined, and if the document contains the noise word, the relationship between the document and the research theme is determined by referring to the noise word.

24. A research support method according to claim 22, comprising: determining whether a noteworthy expression different from the noise word exists in the document; determining the relationship between the document and the research theme if the document does not contain the noise word; determining the relationship between the document and the research theme by referring to the noise word if the document contains the noise word; and determining the relationship between the document and the research theme by referring to the noise word and the noteworthy expression if the document contains both the noise word and the noteworthy expression.

25. A research support system comprising a processor, wherein the processor extracts multiple documents based on a set research theme, and when a user sequentially reviews the multiple documents, if the document being reviewed contains an unregistered noise word, the processor accepts the unregistered noise word based on user operation on the operation unit, stores the accepted noise word in the computer's storage device, and when multiple documents are reviewed, the system makes it possible to distinguish between the noise word stored in the storage device and other words.

26. The research support system according to claim 25, wherein the processor displays the multiple documents sequentially on a display screen when a user sequentially reviews the multiple documents, and highlights the noise words stored in the storage device on the display screen in the multiple documents displayed sequentially on the display screen.

27. The research support system according to claim 25, wherein the processor automatically reads aloud a plurality of documents in order when a user sequentially reviews a plurality of documents, and switches the method of automatically reading aloud the noise words stored in the storage device and words other than the noise words in the plurality of documents that are automatically read aloud in order.

28. The research support system according to claim 25, wherein the processor extracts multiple patent documents as the multiple documents.

29. The investigation support system according to claim 28, wherein the processor makes it possible to distinguish between the noise word and phrases other than the noise word, at least within the claims of the patent document.

30. The research support system according to claim 28, wherein the processor enables the identification of a first patent document in which the noise word is not included in the claims and a second patent document in which the noise word is included in the claims.

31. The research support system according to claim 28, wherein the processor assigns identification information to at least one of the first patent document and the second patent document that distinguishes between a first patent document in which the noise word is not included in the claims and a second patent document in which the noise word is included in the claims, and the identification information is information that can be used when rearranging the first patent document and the second patent document in a plurality of patent documents, or when selecting only the first patent document from a plurality of patent documents as a patent document for peer review.

32. The research support system according to claim 25, wherein the processor displays an input field for inputting the noise word on a display screen that sequentially displays a plurality of the documents, and accepts the noise word to be stored from the input field.

33. The investigation support system according to claim 32, wherein the processor accepts editing of the noise word.

34. The investigation support system according to claim 32, wherein the processor causes the display screen to display a message prompting the input of noise words.

35. The research support system according to claim 25, wherein a display screen that sequentially displays a plurality of the aforementioned documents is provided with an input reception field, the operation unit includes a keyboard, and the keyboard is assigned a first shortcut key that enables the input of noise words into the input reception field, and when the first shortcut key is operated, moves the cursor on the display screen to a predetermined position in the input reception field.

36. The investigation support system according to claim 35, wherein the processor receives the noise word to be stored based on keyboard input, voice input, or copy and paste or drag and drop into the input reception field.

37. The investigation support system according to claim 25, wherein the processor receives the noise words to be stored based on voice input.

38. The display screen is provided with a first registration soft button for storing the noise word entered in the input reception field in the storage device, the operation unit includes a keyboard, and the keyboard is assigned a second shortcut key for moving the cursor over the first soft button, the investigation support system according to claim 35 or 36.

39. The investigation support system according to any one of claims 25 to 37, wherein the processor determines whether or not a noteworthy expression different from the noise word exists in the document, and if the noteworthy expression exists, makes it possible to distinguish between the noteworthy expression and other words or phrases.

40. The research support system according to any one of claims 25 to 37, wherein the processor determines whether a noteworthy expression different from the noise word exists in the document, and if the noteworthy expression exists in the vicinity of the noise word, reduces the distinguishability between the noise word and other words.

41. The search support system according to claim 39, wherein the cautionary expression includes at least one of an alternative statement, a conditional statement, a conditional branch, and a negative expression.

42. The research support system according to any one of claims 25 to 37, wherein the processor receives whether or not the document has been peer-reviewed, and provides each document with at least a first bookmark indicating whether or not the document has been peer-reviewed.

43. The research support system according to claim 42, wherein the display screen that displays a plurality of the aforementioned documents in order is provided with a second soft button for receiving input that the peer review has been completed, and the processor receives whether or not the peer review has been performed based on the user operation of the second soft button.

44. The research support system according to claim 42, wherein the processor accepts that a peer review has been conducted on the document if the document contains an unregistered noise word and the unregistered noise word is stored in the storage device.

45. A display screen that sequentially displays a plurality of the aforementioned documents is provided with two or more third soft buttons that indicate whether or not there is a relationship between the documents and the research theme, or the degree of that relationship, and the processor receives a determination result indicating whether or not there is a relationship or the degree of that relationship based on the user operation of the third soft buttons, and assigns a second bookmark indicating the determination result of the relationship for each of the documents, the research support system according to claim 42.

Images