TWI660278B - A keyword suggestion method for patent search - Google Patents

Abstract

A keyword suggestion method for patent search, including a statistical step of patents of search results, An integrated statistical step of a target search result, a sorting step of selecting a single word appearance count result, and a generating step of a suggested keyword. The statistical step of the search result patent is obtained from a plurality of search result patents containing one or more optional words. One selects a single word group, and then obtains the counted number of words. The integration and statistical steps of the target search result are integrated by the counted number of words, and the integrated result of the counted number of words. The result sorting step is to sort the word count occurrence integration result to obtain a word count occurrence ranking result, and the suggested keyword generation step is to generate a suggested keyword based on the word count occurrence ranking result.

Description

Keyword suggestion method for patent search

The invention relates to a patent search method, in particular to a keyword suggestion method for patent search.

A patent search is performed by a user entering a keyword against a patent database. The keyword may consist of a single keyword or a group of sub-keywords. The keyword is used as a search search condition for the patent data. The library conducts a patent search to obtain a plurality of patent specifications for users to browse and select. The user can perform various patent applications after selection, including search of prior art, patent analysis, and the like. Excellent patent search and search can ensure that: perfect patent layout, accurate patent application and complete patent information collection are extremely important for patent professionals.

However, the conventional patent search system does not have a "logical keyword suggestion mechanism". In the conventional keyword suggestion, the logical method is to record the keyword history of previous users in advance. This conventional method of keyword suggestion has a significant potential problem for the user to retrieve confidentiality leaks. Based on the different needs of each user and the correctness of the keywords themselves, there are doubts. Therefore, in the conventional keyword suggestion methods, it often happens that the search keywords are not accurate enough to get inaccurate. The inaccuracy of patent search results includes: (1) too much and redundant data; or (2) too little and insufficient search results. There are many reasons for the inaccuracy of a keyword, including the use of superior and inferior terms, lack of familiarity with the substantive technical content of the search purpose, and limited language barriers, such as the desire to perform patent searches in other countries but Text Understanding and application is not deep enough and so on. Therefore, in the conventional patent search system, if you want to obtain a complete patent search result, even after repeated keyword exploration and analysis, the obtained patent search result is still subjective to the user, and the entire patent search process Not only is it inefficient, but there are also doubts about accuracy.

Therefore, an object of the present invention is to provide a keyword suggestion method for patent search, which can obtain a suggested keyword for efficient and accurate patent search.

The technical means adopted by the present invention to solve the problems of the conventional technology is to provide a keyword suggestion method for patent search, which regards a plurality of search result patents as a result of a patent search result executed through a search keyword. A proposed keyword is provided for the target search result, and the keyword suggestion method includes: a statistical step of the search result patent, an integrated statistical step of the target search result, a step of selecting a single word count result, and a step of generating a suggested keyword. The statistical step of the search result patent is to treat each search result patent among the plurality of search result patents as a target patent, and for each target patent, a general keyword exclusion table is used to obtain a list containing one or more optional words. An optional word group will be counted according to the number of occurrences of each optional word in the optional word group, and a count result of the number of words corresponding to each optional word will be obtained; the integration and statistical steps of the target search result will be All options for this majority search result patent The word count occurrence results are integrated to obtain a word count occurrence integration result; the selection of a single word count occurrence result is to count the words of all selected words of the plurality of search result patents. The integration result is sorted to obtain a word number appearance ranking result; and the suggested keyword generation step is to generate the suggested keyword according to the word quantity appearance ranking result.

In an embodiment of the present invention, a keyword suggestion method for patent search is provided. In the suggested keyword generating step, a suggestion is generated by selecting at least the highest number of occurrences of a single word in the ranking result of the number of words. A list of keywords to use as suggested keywords.

In one embodiment of the present invention, a keyword suggestion method for patent search is provided. In the suggestion keyword generating step, an exclusion is generated in the order of the number of occurrences of the selected word from the least to the most order. A list of keywords to exclude from the suggested keywords.

According to an embodiment of the present invention, a keyword suggestion method for patent search is provided, wherein the step of integrating statistics of the target search result includes an optional word comparison step, which selects the selected words of the plurality of search result patents and the search. Keywords are compared to each other to form optional words with the same root.

According to an embodiment of the present invention, a keyword search method for patent search is provided, wherein the optional word comparison step includes a radical integration sub-step, and the radical integration sub-step is to select the optional word having the same radical. To be integrated into a root keyword.

In one embodiment of the present invention, a keyword suggestion method for patent search is provided. The universal keyword exclusion list includes prepositions and / or articles and / or nouns that are subject to patent names and / or as a plurality of elements. Connected terms.

In an embodiment of the present invention, a keyword suggestion method for patent search is provided. In the statistical step of the search result patent and / or the integrated statistical step of the target search result, all the selections of the plurality of search result patents are selected. The words are integrated and counted in the place where the patents of the search result appear.

According to an embodiment of the present invention, a keyword suggestion method for patent search is provided, and the appearance position includes: a patent name, a prior art, an embodiment, an abstract, and a scope of patent application.

In an embodiment of the present invention, a keyword suggestion method for patent search is provided. All selected words of the plurality of search result patents have their respective suggested keyword weighting values at the appearance positions of the search result patents.

In an embodiment of the present invention, a keyword suggestion method for patent search is provided, further comprising a re-searching step, which is executed after the suggested keyword generating step, and the re-searching step is executed again with the suggested keyword. The patent search results in a number of patents for re-search results.

According to an embodiment of the present invention, a keyword suggestion method for patent search is provided, and further includes a re-search result patent clustering step. The re-search result patent clustering step is executed after the re-search result patent, the re-search result patent. The clustering step is based on the number of occurrences and the position of the proposed keyword in each of the re-search result patents, and a patent search result point is given to each re-search result patent. The re-search result patent clustering step is based on the patent search. The number of points is obtained to obtain a one-point interval value, and the plurality of re-search result patents are grouped according to the point interval value to obtain a re-search result patent grouping result.

According to an embodiment of the present invention, a keyword suggestion method for patent search is provided, which further includes a step of displaying a search result patent grouping result again, and displaying the search result patent grouping result through a search result patent display element. .

According to an embodiment of the present invention, a keyword suggestion method for patent search is provided, and the appearance position includes: a patent name, a prior art, an embodiment, an abstract, and a scope of patent application.

Through the technical means adopted by the present invention, in the preliminary patent search result, that is, in the search result patent, optional words that appear more frequently or less frequently in all search result patents are obtained, and suggested keywords are generated for use in Searchers conduct further patent searches. Further, preferably, the position of the selected word in the search result patent can be analyzed by analyzing the positions of the selected words, so as to be more accurate. Recognize the importance of using these words as suggested keywords to improve the efficiency and accuracy of patent searches.

A‧‧‧Optional word group

A1, A2, A3, A4, A5, A6, A7‧‧‧Optional words

B‧‧‧ belongs to the root keyword

H1, H2, H3, H4, H5, H6 ‧‧‧ Search result patent

I1, I2, I3, I4, I5

K1‧‧‧Search keywords

K2‧‧‧ suggested keywords

K21, K22, K23, K24, K25 ‧‧‧ sub keywords

S1‧‧‧Statistical steps of search result patents

S2‧‧‧ target search results integration statistical steps

S21‧‧‧Select single word weighting step

S22‧‧‧Select single word comparison step

S221‧‧‧ Alignment sub-step

S222‧‧‧ radical integration sub-step

S3‧‧‧Sorting steps to select counting results

S4‧‧‧ Suggested keyword generation steps

S5‧‧‧Search steps again

S6‧‧‧ Search results again

S7‧‧‧ Search results again Patent grouping result display step

P‧‧‧ Patent Analysis and Retrieval System

P1‧‧‧Search results again Patent display element

FIG. 1 is a flowchart illustrating a keyword suggestion method for a patent search according to an embodiment of the present invention.

FIG. 2 is a screen view showing a keyword suggestion method for patent search in a patent search according to an embodiment of the present invention.

FIG. 3 is a screen view showing a keyword suggestion method for patent search according to an embodiment of the present invention during patent search.

FIG. 4 is a screen diagram showing a keyword suggestion method for patent search according to an embodiment of the present invention during patent search.

FIG. 5 is a screen diagram showing a keyword suggestion method for patent search according to an embodiment of the present invention during patent search.

FIG. 6 is a screen view showing a keyword suggestion method for patent search in a patent search according to an embodiment of the present invention.

FIG. 7 is a screen view showing a keyword suggestion method for patent search in a patent search according to an embodiment of the present invention.

FIG. 8 is a screen view showing a keyword suggestion method for patent search in a patent search according to an embodiment of the present invention.

FIG. 9 is a screen view showing a keyword suggestion method for patent search according to an embodiment of the present invention during patent search.

FIG. 10 is a screen diagram showing a keyword suggestion method for patent search according to an embodiment of the present invention during patent search.

FIG. 11 is a screen view showing a keyword suggestion method for patent search according to an embodiment of the present invention during patent search.

FIG. 12 is a screen diagram showing a keyword suggestion method for patent search in a patent search according to an embodiment of the present invention.

FIG. 13 is a screen diagram showing a keyword suggestion method for patent search in a patent search according to an embodiment of the present invention.

FIG. 14 is a screen view showing a keyword suggestion method for patent search in a patent search according to an embodiment of the present invention.

Embodiments of the present invention will be described below with reference to Figs. 1-14. This description is not intended to limit the embodiment of the present invention, but is an example of the embodiment of the present invention.

As shown in Figs. 1-14, a keyword search method for a patent search according to an embodiment of the present invention is for a patent search result (including a plurality of search result patents H1) executed with a search keyword K1. , H2, H3, H4, H5, H6) as a target search result and provide a suggested keyword K2. The keyword suggestion method of the patent search includes: a statistical step S1 of the search result patent, an integrated statistical step S2 of the target search result, a step S3 of selecting a single-word occurrence count result, a step S4 of suggested keyword generation, a search step S5 again, and again The search result patent grouping step S6 and the search result patent grouping result display step S7.

As shown in FIG. 1, in this embodiment, the statistical result step S1 of the search result patent, the integrated statistical step S2 of the target search result, the selective word appearance count result ranking step S3, and the suggested keyword generation step S4 , The re-searching step S5, the re-searching result patent grouping step S6 And the re-search result patent grouping result display step S7 is executed sequentially. In addition, the keyword suggestion method for patent retrieval of the present invention is mainly implemented in a patent analysis retrieval system P, and is mainly applied to English patent retrieval. Therefore, the examples shown below are presented in English.

As shown in FIG. 2, in detail, the search key K1 is generated in response to a search purpose. In this embodiment, the search purpose is "to make the terminal an easy-to-assemble electrical connector", and the search keyword K1 is set to "connector to make the terminal to be an easy-to-assemble electrical connector" according to the search purpose AND terminal AND pad ", where the connector is" electrical connector ", the terminal is" terminal ", and the pad is" pad ". After the search of the search keyword K1, a plurality of search results patents H1, H2, H3, H4, H5, H6, the plurality of search result patents H1, H2, H3, H4, H5, H6 are the majority of patent documents obtained by searching the search key K1.

Please also refer to Figs. 2-4. The statistical step S1 of the search result patent is for each of the plurality of search result patents H1, H2, H3, H4, H5, and H6. , H4, H5, and H6 are granted as a standard patent. For each target patent, an optional word group A containing one or more optional words is obtained according to a common keyword exclusion table. The optional word group A contains each The selection of single words A1, A2, A3, A4, A5, A6, A7. Calculate the number of times according to the number of occurrences of each of the optional words A1, A2, A3, A4, A5, A6, A7 in the optional word group A, and obtain the corresponding corresponding single words A1, A2, A3, A4, A5. A6, A6, A7 count results.

As shown in FIG. 3, in this embodiment, the search keyword K1 (that is, "connector AND terminal AND pad") is searched and the plurality of search results are patented H1, H2, H3, H4, H5, and H6. The number of words in the Chinese classics statistics obtained `` connection '', `` connector '', `` connected '', `` alignment '', `` pin '', `` terminal '', and `` plug ( Plug) "appears more frequently, and these words are the majority of the optional words A1, A2, A3, A4, A5, A6, A7, the plurality of selected words A1, A2, A3, A4, A5, A6, A7 together form the selected word group A.

As shown in Figure 4, the number of times that the plurality of selected words A1, A2, A3, A4, A5, A6, A7 appear in the plurality of search result patents H1, H2, H3, H4, H5, and H6, respectively, is The word count results. In this embodiment, the selected word "A1" is calculated by the statistical result step S1 of the search result patent and is obtained 160 times in the plurality of search result patents H1, H2, H3, H4, H5, and H6; The "connector" word A2 was calculated after the statistical step S1 of the search result patent and was found in the majority of the search result patents H1, H2, H3, H4, H5, and H6; a total of 230 times; "connected" this optional word A3 is calculated after the statistical step S1 of the search result patent and is found in the majority of the search result patents H1, H2, H3, H4, H5, H6 a total of 368 times; "alignment" This optional word A4 was used in the search result patent The statistical step S1 is calculated to obtain 150 times appearing in the majority of the search result patents H1, H2, H3, H4, H5, and H6; "pin", the selected word A5, is calculated through the statistical step S1 of the search result patent. Later, it was obtained 190 times in the majority of the search result patents H1, H2, H3, H4, H5, and H6; "terminal" is the selected word A6, which is calculated after the statistical step S1 of the search result patent. Appeared 280 times in the majority of the search result patents H1, H2, H3, H4, H5, H6; and "plug" this optional word A7 was calculated in the statistical step S1 of the search result patent and was found in the majority The search results were 70 times for patents H1, H2, H3, H4, H5, and H6.

Continuing to refer to FIG. 4, preferably, the statistical step S1 of the search result patent may further further select all the selected words A1, A2, A3, H1, H2, H3, H4, H5, and H6 of the plurality of search result patents. A4, A5, A6, and A7 are integrated in the "appearance position" of the search result patents H1, H2, H3, H4, H5, and H6. Specifically, the search results of the patents H1, H2, H3, H4, H5, and H6 include the patent name, prior art, implementation, abstract, and scope of patent applications. Following the above, in this embodiment, the word "A1" is selected according to the statistics of the patents of the search result. Step S1 is calculated to obtain the patent names appearing in the plurality of search result patents H1, H2, H3, H4, H5, H6 19 times, 33 times of prior art, 56 times of embodiment, 20 times of abstract, and patent application The range is 32 times; "connector" This selected word A2 is calculated through the statistical step S1 of the search result patent, and the patent names 32 appearing in the plurality of search result patents H1, H2, H3, H4, H5, and H6 are obtained. Times, prior art 50 times, implementations 80 times, abstracts 25 times and patent application scopes 43 times; "connected" This selected word A3 was calculated after the statistical step S1 of the search result patent and appeared in the majority of searches Results The patent names in the patents H1, H2, H3, H4, H5, and H6 were 50 times, the prior art 75 times, the implementation 121 times, the abstract 44 times, and the scope of patent applications 78 times; "alignment" this optional word A4 was subjected to the search The statistical step S1 of the result patent is calculated to obtain the patent names appearing in the plurality of search result patents H1, H2, H3, H4, H5, H6 10 times, 19 times in the prior art, and Embodiment 45. , Abstract 32 times and patent application scope 44 times; "pin" this selected word A5 is calculated by statistical step S1 of the search result patent and obtained in the majority of the search result patents H1, H2, H3, H4, H5 , H6 patent names 28 times, prior art 30 times, implementation 55 times, abstracts 40 times and patent application scope 37 times; "terminal" This selected word A6 is calculated after calculating the statistical step S1 of the search result patent. The names of the patents appearing in the plurality of search result patents H1, H2, H3, H4, H5, H6 32 times, prior art 26 times, embodiments 80 times, abstracts 54 times, and patent applications 88 times; and "plug "The selected word A7 was calculated by the statistical step S1 of the search result patent and obtained the patent names appearing in the majority of the search result patents H1, H2, H3, H4, H5, and H6 5 times, and the prior art 17 times. , 21 times in the embodiment, 12 times in the abstract, and 15 times in the scope of patent application. Alternatively, in other embodiments, the selected statistics of the plurality of search result patents may be integrated and counted in the position of the plurality of search result patents.

In detail, the general keyword exclusion table is used to filter out words that are less likely to appear in the search. The words that are less likely to appear are not meaningful to patent searches. The general keyword exclusion The keywords in the table include: prepositions and / or articles and / or nouns that are subject to patent names. A preposition is a vocabulary or affix used to express the grammatical function of a word in grammar. Prepositions in English include "at", "in", "on", "from", "by", "for", and "of". Taking the English sentence "A stopper which abuts on the surfaces of the first and second substrates." As an example, "on" in the English sentence is a preposition. Articles, mainly before or after nouns or noun phrases, are mainly words that restrict nouns in sentences. Similarly, take the English sentence "A stopper which abuts on the surfaces of the first and second substrates." The term of the patent title refers to the patent name of the patent of the search result. For example, in this embodiment, the patent name of the search result patent H1 is "Electric connector device", where "device" is the noun of the patent name.

As shown in Figures 1 and 5, the integrated statistical step S2 of the target search result is all the selected words A1, A2, A3, A4, A5 of the plurality of search result patents H1, H2, H3, H4, H5, and H6. , A6, A7 The word count appears to be integrated, and a word count appears to be integrated. The integrated statistical step S2 of the target search result includes a selected word weighting step S21 and a selected word comparison step S22.

Please continue to refer to FIG. 5. The optional word weighting step S21 refers to all optional words A1, A2, A3, A4, A5, A6, A7 of the plurality of search result patents H1, H2, H3, H4, H5, and H6. The number of occurrences of patents H1, H2, H3, H4, H5, H6 appearing in the search results is weighted. In detail, all the selected words A1, A2, A3, A4, A5, A6, A7 of the plurality of search result patents H1, H2, H3, H4, H5, H6 appear in the search result patents H1, H2, H3 , H4, H5, H6 have their own suggested keyword weights. In this embodiment, the suggested keywords appearing in the "Patent Application Scope" have a weighted value of 5 points, the suggested keywords appearing in the "Abstract" have a weighted value of 4 points, and the suggested keywords appearing in "Patent Name" have a weighted value. A value of 3 points, a suggested keyword appearing in "Embodiment" has a weighted value of 2 points, and a suggested keyword appearing in "prior art" has a weight It's worth 1 point. Of course, the invention is not limited to this. In other embodiments, the weights of the suggested keywords in the patent application scope, abstract, patent name, implementation mode, and prior art positions in most of the search result patents may be different according to circumstances, such as search purposes. As mentioned above, the "connection" is obtained through the statistical step S1 of the search result patent. This optional word A1 appears 32 times in the patent application scope of the majority of the search result patents H1, H2, H3, H4, H5, and H6. Therefore, 160 points are obtained by multiplying 5 points by 32 times, appearing in the abstract of the majority search results patents H1, H2, H3, H4, H5, and H6 20 times. Therefore, multiplying 4 points by 20 times to obtain 80 points; appearing in The majority search result patents H1, H2, H3, H4, H5, H6 have 19 patent names, so multiplying 3 points by 19 times to get 57 points; appeared in the majority search result patents H1, H2, H3, H4 , H5, H6 in the implementation of 56 times, thus multiplying 2 points by 56 times to get 112 points; and the prior art appeared in the majority of the search results patents H1, H2, H3, H4, H5, H6 33 times, so Multiply 33 times by 1 point to get 33 points. Finally, the points are added up to obtain a total of 442 points for the connection word "A1", which means that the word count appears as a result of integration. By analogy, the word "connector" weighted by the optional word weighting step S21 is counted and the integration result is 621 points; "connected" the selected word A3 is weighted by the optional word weighting step S21 The number of counted integration results is 1033 points; the "alignment" of this selected word A4 is weighted by the selected word weighting step S21 and the number of counted integration results is 487 points; "pin" this selected word A5 is weighted by the selected word The number of words after weighting in step S21 is counted and the integration result is 569 points; the word "terminal" is selected by word A6, and the number of words after weighting in step S21 is counted and the result of integration is 938 points; and "plug" is selected The word count of the single word A7 after the weighting of the selected single word weighting step S21 appears, and the integration result is 197 points. Preferably, if an optional word appears in the scope of patent application and the abstract at the same time, it means that the optional word has a high probability to exist in the first item of the claim in the scope of patent application, and has high importance. Choose a single word to give a higher weight to the suggested keywords. In the keyword suggestion method of the patent search of the present invention, the number of words in a single word is selected and scores are counted and integrated. The higher the number, the greater the number of occurrences and the number of occurrences at key positions. Therefore, it indicates that the selected word may have a higher critical importance.

Please refer to FIGS. 1-3, 6, and 7 at the same time. The optional word comparison step S22 is performed after the optional word weighting step S21, and includes a comparison sub-step S221 and a radical integration sub-step S222. The matching sub-step S221 is to cross the search result patents H1, H2, H3, H4, H5, H6 with the selected words A1, A2, A3, A4, A5, A6, A7 and the search key K1. Compare to get optional words with the same root. Following the above, the optional word comparison step S22 can select the selected words A1, A2, A3, A4, A5, A6, A7 and the search keywords in the plurality of search result patents H1, H2, H3, H4, H5, and H6. K1 compared them and found that "connection", "connector", and "connected" had the same radical, that is, "connect", and were used as optional words. After selection, the optional words A1 and A2 of "connection" and "connector" with similar meaning (electrical connector) can be selected, and the two optional words A1 and A2 are weighted by the optional word in step S21 The counted integration result can be added, that is, the word count of "connection" appears in the counted integration result of 442 points, and the word count of the "connector" appears in the counted integration result of 621 points, and another word count appears in the counted integration. The result is 1063 points. In this way, optional words that have similar meanings (for example, "connect" in this embodiment) but are slightly different in spelling can be integrated. Of course, the present invention does not limit the optional word comparison step S22 to be performed after the optional word weighting step S21. In other embodiments, the optional word comparison step may be performed before the optional word weighting step. That is to say, the keyword suggestion method of the patent search of the present invention can first execute the optional word comparison step to compare the optional words with similar meanings and spelling, and then make these optional words perform the optional word weighting step to The result of word count integration was obtained.

As shown in FIG. 8, the radical integration sub-step S222 is to integrate the selected word having the same radical into a corresponding radical keyword. Following the above "connection" and "connector" The example of the two selected words A1 and A2. The two selected words A1 and A2 can be integrated into the root keyword “connect” by the root integration sub-step S222.

Please refer to Figures 2-3, 5, 7, 9a, and 9b at the same time. The optional word appears in the counting result ranking step. S3 is the selection of the multiple search result patents H1, H2, H3, H4, H5, and H6. The word counts of single words A1, A2, A3, A4, A5, A6, and A7 are counted and the results are sorted to get a single word count. As shown in FIG. 9a, in this embodiment, the result of sorting the number of words is sorted according to the total score of each selected single word A1, A2, A3, A4, A5, A6, A7. Optionally, as shown in FIG. 9b, the word number appearance ordering result may be a counted integration result according to the word number obtained by performing the optional word comparison step S22, that is, the two optional words A1, The word count of A2 "connection" and "connector" is calculated by the result of counting and integration.

As shown in FIG. 10, the suggested keyword generating step S4 is to generate the suggested keyword K2 according to the ranking result of the number of words. In detail, the suggested keyword generation step S4 is to generate a list of suggested keywords by selecting the single words A1, A2, A3, A4, A5, A6, and A7 in the order of appearance of the number of words. As the suggested keyword K2. In this embodiment, the suggested keyword K2 includes the optional words A1, A2, A4, A5, A6 (that is, "connection", "connector", "alignment", "pin", "terminal"). In addition, the combination of the selected words A1, A2, A4, A5, A6 is: "pin" and "terminal" and "alignment" and ("connection" or "connector"). Optionally, the above-mentioned belonging keyword B, that is, "connect", can also be used as one of the suggestions for the keyword K2. Preferably, in other embodiments, the step of generating the suggested keywords may include generating a list of excluded keywords from the order of the occurrences of the selected words in the sorted result of the number of words from as few as possible to exclude from the suggested keywords. That is, the suggested keyword generation step may delete the selected word with the lowest score in the ranking result of word counts instead of the suggested keyword.

As shown in FIGS. 1 and 11, the re-searching step S5 is performed after the suggesting keyword generating step S4. The re-searching step S5 is to perform a re-patent search with the suggested keyword K2 to obtain a plurality of re-search results patents I1, I2, I3, I4, I5. In detail, in this embodiment, the suggested keyword K2 includes five sub-keywords K21, K22, K23, K24, and K25 of "pin", "terminal", "alignment", "connector", and "connection", and The five sub-keywords K21, K22, K23, K24, and K25 are the suggested keywords K2 in a combination of ("pin" and "terminal" and "alignment" and ("connector" or "connection").

As shown in Figures 1 and 12, the re-search result patent clustering step S6 is executed after the re-search result patent clustering step S6 is based on the sub-keywords K21, K22, K23, K24, K25 (ie: "pin", "terminal", "alignment", "connector", and "connection") in each of the re-search result patents, the number of occurrences of I1, I2, I3, I4, I5, and Appearance points are given to each of the re-search results patents I1, I2, I3, I4, I5-a patent search result point. In detail, the appearance position includes the patent name, prior art, implementation, abstract, and scope of patent application. In this embodiment, each of the re-retrieval result patents I1, I2, I3, I4, and I5 each has a keyword weight value for the patent application scope, abstract, patent name, implementation, and prior art. The keyword weight of the patent application range is 5 points, the keyword weight of the abstract is 4 points, the keyword weight of the patent name is 3 points, the keyword weight of the embodiment is 2 points, and the keyword weight of the prior art The value is 1 point. Of course, the invention is not limited to this. In other embodiments, the scope of the patent application, abstract, patent name, embodiment, and keyword weighting of the prior art in the search result patent may be different according to circumstances.

Please continue to refer to FIG. 12. Taking the sub-keyword “K21” as an example, the scope, abstract, patent name, implementation mode, and prior art of the search result Patent I1 appear 10, 15 times, Once, 30 times and 5 times, a total of 61 times. The number of times the patent application scope, abstract, patent name, and implementation mode respectively appear in the re-search result patent I1 are multiplied by the corresponding Keyword weighted value. For example: the weight of the suggested keywords for the scope of patent applications is 5 points, so multiplying the number of occurrences by 10 times to get 50 points; the keyword weight of the abstract is 4 points, so multiplying the number of occurrences by 15 times to get 60 points, according to And so on. And the sub-keyword K21 of "pin" in the re-search result patent I1, the patent application scope, abstract, patent name, implementation mode and the weighted score of the prior art are added to obtain the "pin" sub-keyword K21 in The score of the search result patent I1 was 178 points. Similarly, after weighted calculation, the sub-keyword K21 of "pin" has scores of 313 points, 179 points, 100 points, and 400 points in the re-search result patents I2, I3, I4, and I5, respectively.

As shown in Figure 13, according to the above calculation method, the scores of each of the sub-keywords K22, K23, and K24 in the re-search result patent I1I2, I3, I4, and I5 are calculated and added to obtain the value of each The search results of the patents I1, I2, I3, I4, and I5 were 738 points, 513 points, 509 points, 232 points, and 1098 points.

As shown in FIG. 14, the re-search result patent clustering step S6 is based on the points of the patent search result, that is, 738 points, 513 points, 509 points, 232 points, and 1098 points to obtain a one-point interval value of 200 to 500. , 500 ~ 800 and 800 ~ 1100. The re-search result patent clustering step S6 is to group the plurality of re-search result patents I1, I2, I3, I4, and I5 to obtain a re-search based on the points interval values of 200 to 500, 500 to 800, and 800 to 1100. Results patent clustering results. The re-search result patent grouping result display step S7 is to display the re-search result patent grouping result through a re-search result patent display element P1, and the re-search result patent display element P1 includes a display.

The above descriptions and descriptions are merely illustrations of the preferred embodiments of the present invention. Those with ordinary knowledge of this technology may make other modifications based on the scope of the patent application defined below and the above description, but these modifications should still be made. It is the spirit of the present invention and is within the scope of the present invention.

Claims (13)

A keyword suggestion method for patent search is to provide a suggestion keyword for a plurality of search result patents of a patent search result executed through a search keyword as a target search result. The keyword suggestion method Including: the statistical step of the search result patent is to treat each search result patent among the plurality of search result patents as a target patent, and obtain a single word from each target patent by a processor, and assign all the single words One or more optional words are obtained by excluding a general keyword according to a general keyword exclusion table, and the obtained one or more optional words constitute an optional word group, and the processor includes the one or more optional words. A plurality of the optional words of the optional word are stored in a memory; an integrated statistical step of the target search result is taken out from the memory by the processor, the option containing one or more of the optional words Word groups, and the processor will count occurrences of all selected words for the plurality of search result patents Calculate, and get the count-by-word integration result corresponding to each word of the selected word, the processor stores the count-by-word integration result of the word count in the memory; step of selecting the word-by-word count result , By using the processor to fetch the word counts from the memory, a counted integration result appears, and the processor sorts the word counted integration results of all selected words of the plurality of search result patents. To obtain a word number appearance ordering result, the processor stores the word number appearance ordering result in the memory; and propose a keyword generation step to retrieve the word number appearance order from the memory by the processor. As a result, the processor generates the suggested keyword according to the sorted result of the word count, and the processor stores the suggested keyword in the memory. For example, the method for keyword suggestion in the scope of patent application, wherein in the step of generating the suggested keyword, the processor generates a suggestion key in at least the order of the number of occurrences of the selected word in the ranking result of the number of words. The word list is stored in the memory as the suggested keyword. For example, the method for keyword suggestion in the scope of patent application, wherein in the step of generating the suggested keyword, the processor generates an exclusion key by selecting the number of occurrences of the selected word in the sorting result of the number of words from the least to the most order. The word list is stored in the memory to exclude from the suggested keywords. For example, the keyword suggestion method of the first scope of the patent application, wherein the step of integrating statistics of the target search result includes an optional word comparison step, the processor selects the selected words of the plurality of search result patents and the search keyword. They are compared with each other to obtain an optional word having the same radical and are stored in the memory. For example, the method for keyword suggestion in the fourth item of the patent application, wherein the optional word comparison step includes a radical integration sub-step, and the processor integrates the optional words with the same radical by the processor. A root keyword is stored in the memory. For example, the keyword suggestion method of the first scope of patent application, wherein the general keyword exclusion table includes prepositions and / or articles and / or nouns that are subject to patent names and / or serves as a connection relationship between a plurality of elements. Words. For example, the keyword suggestion method of the first patent application scope, wherein in the statistical step of the search result patent and / or the integrated statistical step of the target search result, all optional words of the plurality of search result patents are processed by the processor. Integrate statistics at the appearance position of the search result patent and store it in the memory. For example, the keyword suggestion method for item 7 of the scope of patent application, where the appearance position includes: patent name, prior art, implementation, abstract, and scope of patent application. For example, the method for keyword suggestion in the seventh scope of the patent application, wherein all the selected words of the plurality of search result patents have their respective suggested keyword weighting values at the appearance positions of the search result patents. For example, the method for keyword suggestion in the first item of the patent application scope further includes a re-searching step, which is executed after the suggestion keyword generation step, and the re-searching step is to retrieve the suggested keyword from the memory by the processor. , And execute a re-patent search with the suggested keywords to obtain a plurality of re-search result patents, and store the plurality of re-search result patents in the memory. For example, the method for keyword suggestion of the 10th patent application scope further includes a re-search result patent clustering step. The re-search result patent clustering step is executed. After the re-search step, the re-search result patent clustering step is processed by the process. The device retrieves the suggested keywords from the memory to give each re-search result patent a patent search result point according to the number of occurrences and the position of the suggested keyword in each re-search result patent, and The points of the patent search result are stored in the memory, and the step of re-searching the patent clustering step is that the processor retrieves the points of the patent search result from the memory to obtain a point according to the points of the patent search result Count the interval value, and store the point interval value in the memory, the processor fetches the point interval value from the memory, and retrieves the majority of the result patents based on the point interval value Grouping to get a re-search result patent grouping result, and re-searching the patent grouping result To store in the memory. For example, the keyword suggestion method for item 11 of the patent application scope further includes a re-search result patent clustering result display step, and the processor displays the re-search result patent clustering result through a re-search result patent display element. For example, the keyword suggestion method for item 11 of the scope of patent application, where the appearance position includes: patent name, prior art, implementation, abstract, and scope of patent application.