Patent document retrieval method, electronic device and computer readable storage medium
By setting multiple search conditions in patent literature search and using semantic search methods, the patent literature is deduplicated and sorted, and the problem of low search efficiency and accuracy in the existing technology is solved, and efficient and accurate patents are achieved Literature search results.
Patent Information
- Application Number
- CN202011068820.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2020-09-30
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2040-09-30
AI Technical Summary
The existing patent literature search methods are difficult to quickly find useful documents in massive literature libraries, and the search results are low in accuracy, and the same literature cannot be effectively deduplicated, which increases the screening workload.
By setting multiple search conditions, obtaining patent documents using semantic search, Boolean search, and homoethnic relationships, and digitally labeling the same documents, deduplication processing and sorting, improving search efficiency and accuracy.
The number of duplicate documents in the search results is reduced, the search efficiency and accuracy is improved, and the similarity with the target patent documents is intuitively displayed through digital labels, which facilitates users to quickly filter.
Smart Images

Figure CN112182030B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of information retrieval technology, and in particular, to a patent document retrieval method, an electronic device using the patent document retrieval method, and a computer-readable storage medium for executing the patent document retrieval method. Background Art
[0002] Patent documents contain a large amount of information on technology, economy, law, etc., and have become a key factor in the advancement of intellectual property strategy. Some of the most advanced and valuable inventions are always first published in patent documents. The novelty of the content makes patent documents the easiest and fastest tool for finding the latest progress in a certain field.
[0003] Patent documents provide a rich knowledge treasure house and have become the main carrier of scientific and technological progress and innovation. The acquisition, creation, application and management of intellectual property rights are inseparable from the search for similar patent documents. For example, when examiners examine whether a patent meets the three characteristics (novelty, creativity and practicality), they need to retrieve the prior art including patent documents, or when invalidating a patent, they need to search for patent documents similar to the invalidated patent, or when technicians search for patent documents similar to R&D technology, etc. Therefore, how to quickly find useful patent documents in a massive patent document library is a greater challenge.
[0004] Existing patent searches usually obtain patent data sets by constantly changing search terms. However, the same patent documents exist between the patent data sets obtained by different search terms, and these same patent documents are not deduplicated, which increases the screening workload and reduces the accuracy of the search results.
[0005] Therefore, there is an urgent need to provide a new patent document retrieval method. Summary of the invention
[0006] The present invention aims to provide a patent document retrieval method with high retrieval efficiency and high retrieval result accuracy.
[0007] At the same time, an electronic device using the above patent document retrieval method is also provided.
[0008] Furthermore, the present invention also provides a computer-readable storage medium for executing the patent document retrieval method.
[0009] A patent document retrieval method includes the following steps: setting a first search condition to perform a search and obtaining a first patent document set, wherein the first document set includes multiple patent documents; setting a second search condition to perform a search and obtaining a second patent document set, wherein the second document set includes multiple patent documents; matching multiple patent documents in the second patent document set with multiple patent documents in the first patent document set to determine whether they are identical; if a patent document in the second patent document set is identical to a patent document in the first patent document set, adding a digital tag to the identical patent document.
[0010] Furthermore, the second search condition includes at least one of semantic search, Boolean search, keyword search, family relationship, citation relationship, IPC classification number, applicant, inventor, CPC classification number and cited relationship.
[0011] Furthermore, the second search condition also includes region, country, and patent document database limitations.
[0012] Furthermore, the method further includes: changing the second search condition in the second search condition, repeatedly performing the search, and acquiring a plurality of the second patent document sets.
[0013] Furthermore, the method also includes: setting a third search condition to perform a search to obtain a third patent document set, wherein the third patent document set includes multiple patent documents, and if the patent documents in the third patent document set are the same as the patent documents in the first patent document set, adding a digital tag to the same patent document.
[0014] Furthermore, when adding a digital tag to the same first patent document, the number of times the patent documents in the first patent document set are identical to the patent documents in other patent document sets is calculated, and the digital tag value is equal to the numerical value obtained by subtracting 1 from the identical number of times.
[0015] Furthermore, the method also includes: sorting the patent documents in the first patent document set according to the digital label; selecting patent documents according to the sorting result, wherein the larger the value of the digital label is, the higher the similarity between the patent document and the target patent document is.
[0016] Furthermore, the method also includes: adding color marks to the browsed patent documents under the data nodes.
[0017] Furthermore, the method also includes: adding color marks to the browsed patent documents.
[0018] Furthermore, the first search condition is related to the second search condition.
[0019] Furthermore, the first search condition is a target patent document number, and the second search condition is a semantic search corresponding to the claims of the target patent document number obtained according to the first search condition.
[0020] Furthermore, the method further includes performing an AND operation on the first patent document set and the second patent document set, merging patent documents identical to those in the first patent document set, and marking digital tags for identical patent documents.
[0021] An electronic device includes a processor and a memory. When the memory stores an execution command, the processor and the memory communicate with each other, and the processor executes the patent document retrieval method as described above.
[0022] A computer-readable storage medium for executing a patent document retrieval method includes a computer-readable storage medium storing a program code, wherein the program code includes instructions for executing the patent document retrieval method as described above.
[0023] Compared with the prior art, in the patent document retrieval method of the present invention, by performing multiple searches on the target patent document, such as semantic search, and deduplicating the patent data set obtained by each semantic search and the patent documents in the data set updated last time and digitally marking the same patent documents, the number of patent documents retrieved is reduced, thereby reducing the workload of screening, and the patent document closest to the target patent document can be intuitively obtained through the digital labels on the patent documents, with high retrieval accuracy and efficiency.
[0024] More importantly, during the search process, two different search conditions can be set, but the two related search conditions are searched independently. In this way, the two search results can be quickly calculated and deduplicated. The patent documents obtained by repeated searches are digitally labeled to indicate the number of hits and improve their importance. The important patent documents in the search results are displayed visually, which is convenient for users to quickly screen.
[0025] Furthermore, it is possible to set related but different search conditions an unlimited number of times to obtain multiple independent patent document sets, perform AND operations on the results obtained from multiple searches, mark the number of hits, and sort them according to the number of hits, and display the patent documents with more hits at the top to facilitate user browsing. BRIEF DESCRIPTION OF THE DRAWINGS
[0026] Figure 1 It is a schematic flow chart of the patent document retrieval method disclosed in the first embodiment of the present invention;
[0027] Figure 2 is a flow chart of a patent document retrieval method disclosed in the second embodiment of the present invention;
[0028] Figure 3 is a flow chart of a patent document retrieval method disclosed in the second embodiment of the present invention;
[0029] Figure 4 It is a structural schematic diagram of an electronic device of the present invention. DETAILED DESCRIPTION
[0030] The technical solutions in the embodiments of the present invention are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0031] The patent document retrieval method of the present invention can be implemented by using an electronic device. The electronic device can be implemented by hardware or software. The electronic device can be integrated in a computer to implement the patent document retrieval method.
[0032] Embodiment 1
[0033] See also Figure 1 , is a flow chart of a patent document search method disclosed in the first embodiment of the present invention. The patent document search method comprises the following steps:
[0034] Step S11, setting a first search condition according to the target patent document to perform a search and obtain a first patent document set;
[0035] In step S11, the first patent document set is a basic search result set obtained by executing a search command, and the basic search result set is a patent document set composed of multiple patent documents. For example, assuming that the publication number of the target patent document is cn1234567, the first search condition can be set to r / cn1234567, that is, use the publication number cn1234567 as the first search condition; enter "r / cn1234567" in the search window and select the Chinese database; then, execute the search command; finally, obtain several Chinese patent publication texts and authorized announcement patents. Of course, it is not limited to using the publication number of the target patent document as the search term. You can also select a certain technical feature in the target patent document as the search term. For example, the search formula is: a / cdma and isd / 2000-2019, which means that the invention name, abstract or claim contains the keyword "cdma", and the patent application text or patent text with a publication date between 2000 and 2019 is used as the search condition. The type of patent document database selected for retrieval is determined according to the country or region to which the target patent document belongs. If the target patent document is a Chinese patent, select the Chinese patent database. If the target patent document is a US patent, select the US patent database. In other words, you can define the search conditions according to the target patent document.
[0036] Step S12, storing each patent document in the first patent document set under a data node;
[0037] Preferably, in step S12, each patent document in the first patent document set is stored under the data node in descending order according to the relevance to the target patent document. The higher the relevance, the closer it is to the target patent document.
[0038] Step S13, setting a second search condition according to the target patent document to perform enhanced search and obtain a second patent document set, wherein the second search condition includes a specific relationship and a restriction condition, wherein the specific relationship is a relationship between documents and the restriction condition is a restriction condition for the document;
[0039] In step S13, the second patent document set is an enhanced search result set obtained by executing the search command, and the enhanced search result set is a patent document set composed of multiple patent documents, wherein the specific relationship includes at least one of Boolean search, semantic search, keyword search, family relationship, citation relationship, IPC classification number, CPC classification number and cited relationship, and the restriction condition includes at least one of region, country, and patent document database type limitation. For example, the second search condition can be set to express that the specific relationship is a family relationship, and the restriction condition is the Chinese English database, that is, enter "r / cn1234567_en and db / ce and fmdb / cnapp" in the search window, and select the Chinese English database, which means: search for Chinese family patents with patent publication number CN1234567 in the Chinese English database; if it is necessary to search for Chinese family patents with patent publication number CN1234567 in the American Chinese database, enter "r / cn1234567_en and db / uc and fmdb / cnapp" in the search window. fmdb / cnapp"; then, execute the search command; finally, obtain several Chinese patent publication texts and authorized announcement patents.
[0040] Step S14, matching each patent document in the second patent document set with each patent document under the data node in turn to determine whether they are the same;
[0041] Step S15, when a patent document in the second patent document set is identical to a patent document under the data node, a digital label is added to the patent document under the data node, and patent documents in the second patent document set that are different from the patent documents under the data node are stored under the data node and after the patent documents under the original data node, so as to update the patent documents under the data node;
[0042] Step S14 and step S15 are performed simultaneously, which essentially allows the computer's CPU to calculate one more time to find the intersection of the same family enhancements in the result set. Each time a patent document under the data node is hit by the enhanced search, the value of the digital label corresponding to the patent document increases by 1. If the document is hit by the enhanced search for the first time, a digital label is added to the patent document. Patent documents in the patent document set obtained by the enhanced search that are different from the patent documents under the data node are stored under the data node and are located after the patent document under the data node after the last update.
[0043] Step S16, sorting the patent documents under the data node according to the digital label;
[0044] Step S17, selecting patent documents according to the sorting result, the larger the value of the digital label is, the higher the similarity between the patent document and the target patent document is.
[0045] At this point, a homology enhanced search is completed.
[0046] During the search process, you can click on any patent document stored under the data node to browse. In order to distinguish whether a patent document has been browsed, the browsed patent document under the data node can be marked with a color.
[0047] Embodiment 2
[0048] See also Figure 2 , is a flow chart of the patent document enhanced retrieval method disclosed in the second embodiment of the present invention. The difference between this embodiment and the first embodiment is that this embodiment is a semantic enhanced retrieval. Specifically, the patent document enhanced retrieval method includes the following steps:
[0049] Step S21, setting a first search condition according to the target patent document to perform a search and obtain a first patent document set;
[0050] In step S21, the first patent document set is a basic search result set obtained by executing a search command, and the basic search result set is a patent document set composed of multiple patent documents. For example, assuming that the publication number of the target patent document is cn1234567, the first search condition can be set to r / cn1234567, that is, use the publication number cn1234567 as the first search condition; enter "r / cn1234567" in the search window and select the Chinese database; then, execute the search command; finally, obtain several Chinese patent publication texts and authorized announcement patents. Of course, it is not limited to using the publication number of the target patent document as the search term. You can also select a certain technical feature in the target patent document as the search term. For example, the search formula is: a / cdma and isd / 2000-2019, which means that the invention name, abstract or claim contains the keyword "cdma", and the patent application text or patent text with a publication date between 2000 and 2019 is used as the search condition. The type of patent document database selected for retrieval is determined according to the country or region to which the target patent document belongs. If the target patent document is a Chinese patent, select the Chinese patent database. If the target patent document is a US patent, select the US patent database. In other words, you can define the search conditions according to the target patent document.
[0051] Step S22, storing each patent document in the first patent document set under a data node;
[0052] Preferably, in step S22, each patent document in the first patent document set is stored under the data node in descending order according to the relevance to the target patent document, and the higher the relevance, the closer it is to the target patent document.
[0053] Step S23, selecting fields in the target patent document to perform semantic enhancement search to obtain a third patent document set;
[0054] In step S23, the third patent document set is also a text data set obtained by executing a set search command, and any field in the target patent document can be selected for semantically enhanced retrieval, and the field can be a claim or a field in a claim, or any field in the specification.
[0055] Step S24, matching each patent document in the third patent document set with each patent document under the data node in turn to determine whether they are the same;
[0056] Step S25, when a patent document in the third patent document set is identical to a patent document under the data node, a digital label is added to the patent document under the data node, and the patent document in the third patent document set that is different from the patent document under the data node is stored under the data node and after the patent document under the original data node, so as to update the patent document under the data node;
[0057] Step S24 and step S25 are performed simultaneously, which is essentially to let the computer's CPU calculate one more time to find the intersection of semantic enhancement in the result set. Each time a patent document under the data node is hit by the enhanced search, the value of the digital label corresponding to the patent document increases by 1. If the document is hit by the enhanced search for the first time, a digital label is added to the patent document. Patent documents in the patent document set obtained by the enhanced search that are different from the patent documents under the data node are stored under the data node and are located after the patent documents under the data node after the last update. Preferably, the initial value of the digital label is 1.
[0058] Step S26, sorting the patent documents under the data node according to the digital label;
[0059] Step S27, selecting patent documents according to the sorting result, the larger the value of the digital label is, the higher the similarity between the patent document and the target patent document is.
[0060] At this point, a semantically enhanced retrieval is completed.
[0061] Embodiment 3
[0062] See also Figure 3, which is a flow chart of the patent document enhanced retrieval method disclosed in the third embodiment of the present invention. The difference between this embodiment and the first and second embodiments is that this embodiment combines the same family enhanced retrieval and the semantic enhanced retrieval. Specifically, the patent document enhanced retrieval method includes the following steps:
[0063] Step S31, setting a first search condition according to the target patent document to perform a search and obtain a first patent document set;
[0064] In this step S31, the first patent document set is a basic search result set obtained by executing a search command, and the basic search result set is a patent document set composed of multiple patent documents. For example, assuming that the publication number of the target patent document is cn1234567, the first search condition can be set to r / cn1234567, that is, use the publication number cn1234567 as the first search condition; enter "r / cn1234567" in the search window and select the Chinese database; then, execute the search command; finally, obtain several Chinese patent publication texts and authorized announcement patents. Of course, it is not limited to using the publication number of the target patent document as the search term. You can also select a certain technical feature in the target patent document as the search term. For example, the search formula is: a / cdma and isd / 2000-2019, which means that the invention name, abstract or claim contains the keyword "cdma", and the patent application text or patent text with a publication date between 2000 and 2019 is used as the search condition. The type of patent document database selected for retrieval is determined according to the country or region to which the target patent document belongs. If the target patent document is a Chinese patent, select the Chinese patent database. If the target patent document is a US patent, select the US patent database. In other words, you can define the search conditions according to the target patent document.
[0065] Step S32, storing each patent document in the first patent document set under a data node;
[0066] Preferably, each patent document in the first patent document set is stored under the data node in descending order according to the relevance to the target patent document. The higher the relevance, the closer it is to the target patent document.
[0067] Step S33, setting a second search condition according to the target patent document to perform enhanced search and obtain a second patent document set, wherein the second search condition includes a specific relationship and a restriction condition, wherein the specific relationship is a relationship between documents and the restriction condition is a restriction condition for the document;
[0068] In step S33, the second patent document set is an enhanced search result set obtained by executing the search command, and the enhanced search result set is a patent document set composed of multiple patent documents, wherein the specific relationship includes at least one of a family relationship, a citation relationship, an IPC classification number or a cited relationship, and the restriction condition includes at least one of a region, a country, and a patent document database limitation. For example, the second search condition can be set to express that the specific relationship is a family relationship, and the restriction condition is the Chinese English database, that is, enter "r / cn1234567_en and db / ce andfmdb / cnapp" in the search window, and select the Chinese English database, which means: search for Chinese family patents with patent publication number CN1234567 in the Chinese English database; if you need to search for Chinese family patents with patent publication number CN1234567 in the American Chinese database, enter "r / cn1234567_en and db / uc and fmdb / cnapp" in the search window; then, execute the search command; finally, obtain several Chinese patent publication texts and authorized announcement patents. Preferably, in this step, the restriction condition in the second search condition may be changed, and step S33 may be repeatedly executed to obtain a plurality of the second patent document sets.
[0069] Step S34, selecting fields in the target patent document to perform semantic enhancement search to obtain a third patent document set;
[0070] In this step, any field in the target patent document can be selected for semantic enhancement search, and the field can be a claim or a field in a claim, or any field in the specification. Of course, the field can also be changed, and the semantic enhancement search can be repeated to obtain multiple sets of the third patent documents, and each time the semantic enhancement search is performed, the fields selected from the target patent document are different. It is worth noting that the order of step S33 and step S34 is not in particular order.
[0071] Step S35, matching each patent document in the second patent document set and the third patent document set with each patent document under the data node in turn to determine whether they are the same;
[0072] Step S36, when a patent document in the second patent document set is identical to a patent document under the data node, a digital label is added to the patent document under the data node, and patent documents in the second patent document set that are different from the patent documents under the data node are stored under the data node and after the patent documents originally under the data node, so as to update the patent documents under the data node;
[0073] Step S37, when a patent document in the third patent document set is identical to a patent document under the data node, a digital label is added to the patent document under the data node, and the patent document in the third patent document set that is different from the patent document under the data node is stored under the data node and after the patent document under the original data node, so as to update the patent document under the data node;
[0074] The order of step S36 and step S37 is not particular, and only one patent document set obtained by enhanced search is selected each time to match with each patent document under the data node after the last update to determine whether they are the same. Each time a patent document under the data node is hit by enhanced search, the value of the digital label corresponding to the patent document increases by 1. If the patent document is hit by enhanced search for the first time, a digital label is added to the patent document. Patent documents in the patent document set obtained by enhanced search that are different from the patent documents under the data node are stored under the data node and are located after the patent documents under the data node after the last update.
[0075] Step S38, sorting the patent documents under the data node according to the digital label;
[0076] Step S39, selecting patent documents according to the sorting result, the larger the value of the digital label is, the higher the similarity between the patent document and the target patent document is.
[0077] At this point, one sibling enhanced retrieval and one semantic enhanced retrieval are completed.
[0078] When adding digital labels to the patent documents under the data node, the initial value of the digital label is preferably 1, and digital labels with different values can also be displayed in different colors to facilitate quick search for patent documents with the same digital label based on color.
[0079] Furthermore, when the patent document under the data node is updated, the name of the data node is updated accordingly, and the name of the data node can be updated through a set template.
[0080] Of course, the enhanced search for the target patent document is not limited to one or two times. If multiple enhanced searches are required, that is, step S33 or step S34 is repeatedly executed, thereby realizing multiple enhanced searches.
[0081] Compared with the prior art, in the patent document enhanced retrieval method of the present invention, by performing multiple semantic enhanced searches on the target patent document, and deduplicating the patent data set obtained by each semantic enhanced search and the patent document under the data node of the previous update and digitally marking the same patent document, the number of patent documents retrieved is reduced, thereby reducing the screening workload, and the patent document closest to the target patent document can be intuitively obtained through the digital labels on the patent document, with high retrieval accuracy and efficiency.
[0082] Furthermore, as a specific application of the patent document enhanced retrieval method described in the present invention, when a patent database needs to be updated, the update of the database is not limited to the update of the database on the timeline, it can also be an operation between two databases, and the update operation instruction can be specifically, for example: it can be to modify data content, delete data content, and add new data content; the update content also includes a root node name identifier, and the root node name identifier can be the name of the tree to which the node to be updated belongs, and the node to be updated is a child node of the tree whose name is the root node name identifier; the update message also includes a keyword identifier of the node to be updated, and the keyword identifier is used to identify the keyword corresponding to the data content for which the update operation specified in the update operation instruction needs to be executed. The electronic device can determine the position of the root node of the node to be updated based on the root node name identifier.
[0083] See also Figure 4 , is a schematic diagram of the structure of an electronic device of the present invention. The electronic device 20 includes a processor 21 and a memory 22. When the memory 22 stores an execution command, the processor 21 and the memory 22 communicate with each other, and the processor 21 executes the execution command so that the electronic device executes the patent document enhanced retrieval method of the present invention.
[0084] Those skilled in the art can understand that all or part of the steps of the above-mentioned patent document enhanced retrieval method embodiment can be completed by hardware related to program instructions, and the above-mentioned program can be stored in a computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiment are executed; and the above-mentioned storage medium includes: ROM, RAM, magnetic disk or optical disk and other media that can store program codes. Therefore, the present invention also provides a computer program product, which includes a computer-readable storage medium storing program codes, and the program code includes instructions for executing the patent document enhanced retrieval method described in the above-mentioned embodiment.
[0085] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.
Claims
1. A patent document retrieval method, characterized in that: The steps include: Setting a first search condition according to the target patent document to be searched to perform a search and obtain a first patent document set, wherein the first patent document set includes a plurality of patent documents; Setting a second search condition according to the target patent document to be searched to perform a search and obtain a second patent document set, wherein the second patent document set includes a plurality of patent documents; Matching the plurality of patent documents in the second patent document set with the plurality of patent documents in the first patent document set to determine whether they are identical; If the patent document in the second patent document set is the same as the patent document in the first patent document set, a digital label is added to the corresponding identical patent document, and the digital label represents the degree of relevance between the patent document and the target patent document; A third search condition is set to perform a search, and a third patent document set is obtained, wherein the third patent document set includes a plurality of patent documents. If a patent document in the third patent document set is identical to a patent document in the first patent document set, a digital tag is added to the identical patent document. When a digital tag is added to the identical patent document, the number of identical times of the patent document in the first patent document set and the patent document in the other patent document set is calculated, and the digital tag value is equal to the number of identical times minus 1, and is consistent with the value obtained by; sorting the patent documents in the first patent document set according to the digital tags; Patent documents are selected according to the ranking results, and the larger the value of the digital label is, the higher the similarity between the patent document and the target patent document is.
2. The patent document retrieval method according to claim 1, characterized in that: The second search condition includes at least one of semantic search, Boolean search, keyword search, family relationship, citation relationship, IPC classification number, applicant, inventor, CPC classification number and cited relationship.
3. The patent document retrieval method according to claim 2, characterized in that: The second search condition also includes region, country, and patent document database limitations.
4. The patent document retrieval method according to claim 1, characterized in that: The method further comprises: The second search condition in the second search condition is changed, and the search is repeatedly performed to obtain a plurality of the patent document collections.
5. The patent document retrieval method according to claim 1, characterized in that: The method further comprises: adding color marks to the browsed patent documents.
6. The patent document retrieval method according to claim 1, characterized in that: The first search condition is related to the second search condition.
7. The patent document retrieval method according to claim 1, characterized in that: The first search condition is a target patent document number, and the second search condition is a semantic search corresponding to the claims of the target patent document number obtained according to the first search condition.
8. The patent document retrieval method according to claim 1, characterized in that: The method also includes performing an AND operation on the first patent document set and the second patent document set, merging patent documents in the second patent document set that are identical to patent documents in the first patent document set, and marking digital tags for identical patent documents.
9. An electronic device, characterized in that: The method comprises a processor, a memory and a computer program stored in the memory and capable of running on the processor, wherein when the computer program is executed by the processor, the steps of the patent document retrieval method as claimed in any one of claims 1 to 8 are implemented.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the patent document retrieval method according to any one of claims 1 to 8 are implemented.
