A method and system for intelligent screening and querying of electronic documents

Abstract

The present invention discloses an electronic document intelligent screening query method and system, which relates to the field of information retrieval, including: generating an electronic document database; generating a query interface; establishing a keyword similarity comparison mechanism; obtaining a keyword package; obtaining a search weight of a target keyword set by a queryer; forming at least one search package; obtaining a search package sequence; obtaining a target electronic document sequence; generating key contents of a common part and key contents of a difference part of a target electronic document; and re-querying using an adjusted similarity allowable coefficient. By setting a mechanism establishment module, a weight acquisition module, a search package formation module, a difference preview module, and a feedback adjustment module, intelligent matching of query keywords can be performed according to the content input by the queryer, enabling the queryer to perform a quick comparison, complete the screening of electronic documents, and autonomously adjust the search accuracy.

Description

Technical Field

[0001] The present invention relates to the field of information retrieval, and in particular to an electronic document intelligent screening and query method and system. Background Art

[0002] Electronic documents refer to texts formed by people in social activities, which are carried by chemical magnetic materials such as computer disks, magnetic disks and optical disks. They rely on computer systems for access and can be transmitted on communication networks. They mainly include electronic documents, electronic letters, electronic reports, electronic drawings, electronic versions of paper text documents, etc.

[0003] When searching for existing electronic documents, the searcher usually selects keywords independently, and the search accuracy is fixed and cannot be adjusted. At the same time, the obtained electronic documents lack intuitive display and comparison, making it difficult for the searcher to quickly identify whether the electronic document is the one he needs. Summary of the invention

[0004] In order to solve the above technical problems, a method and system for intelligent screening and querying of electronic documents are provided. This technical solution solves the problem mentioned in the above background technology that when querying existing electronic documents, the queryer usually selects keywords independently, and the query accuracy is fixed and cannot be adjusted. At the same time, the obtained electronic documents lack intuitive display and comparison, and it is difficult for the queryer to quickly identify whether the electronic document is the document he needs.

[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0006] An intelligent screening and query method for electronic documents, comprising:

[0007] Aggregating at least one electronic document to generate an electronic document database;

[0008] Generate a query interface, the query interface comprising a search box and at least one display box, the at least one display box being arranged in parallel;

[0009] Establish a keyword similarity comparison mechanism, which is determined by the similarity allowable coefficient;

[0010] Generating keywords for electronic documents in the electronic document database to obtain keyword packages, and pairing the electronic documents with the keyword packages;

[0011] When the queryer knows the name of the electronic document, the electronic document name input by the queryer is obtained, and the corresponding electronic document is searched;

[0012] When the inquirer does not know the name of the electronic document, at least one search term input by the inquirer is obtained, and at least one to-be-screened keyword is matched based on a search term and keyword similarity comparison mechanism, and at least one target keyword determined by the inquirer from the at least one to-be-screened keyword is obtained, wherein the to-be-screened keyword and the corresponding search term are similar keywords;

[0013] Obtain the search weight of the target keyword set by the queryer;

[0014] Based on possible combinations of at least one target keyword, forming at least one search package;

[0015] Based on the search weight of the target keyword, the search packages are sorted to obtain a search package sequence, and the search packages are arranged in order at the top of the display box according to the search package sequence;

[0016] Using a keyword similarity comparison mechanism, obtaining at least one electronic document corresponding to the search package as a target electronic document, sorting the target electronic documents based on the search weight of the target keyword to obtain a target electronic document sequence, and displaying the target electronic documents at the bottom of the search package display area in the order of the target electronic document sequence;

[0017] Generate key contents of the common part and key contents of the difference part of the target electronic document, display the key contents of the common part in the search package display area, and display the key contents of the difference part in the display area corresponding to the target electronic document;

[0018] Based on the inquirer's feedback on the query result, the similarity allowance coefficient is adjusted, and the query is re-performed using the adjusted similarity allowance coefficient.

[0019] Preferably, the establishment of a keyword similarity comparison mechanism comprises the following steps:

[0020] Set the similarity tolerance coefficient for the keyword similarity comparison mechanism;

[0021] Obtain two keywords to be compared, respectively as the first keyword to be compared and the second keyword to be compared;

[0022] Acquire at least one first sample document in which a first keyword to be compared appears;

[0023] Replacing the first keyword to be compared in the first sample document with the second keyword to be compared to obtain a second sample document;

[0024] Retrieve the second sample document based on the big data to determine whether there is an electronic document whose overlap with the second sample document is greater than a preset value. If yes, use the second sample document as the target second sample document; if no, do not do anything.

[0025] Divide the number of target second sample documents by the number of first sample documents to obtain the similarity between the first keyword to be compared and the second keyword to be compared;

[0026] When the similarity between the first keyword to be compared and the second keyword to be compared is greater than the similarity allowable coefficient, the first keyword to be compared and the second keyword to be compared are similar keywords.

[0027] Preferably, the step of generating keywords for the electronic documents in the electronic document database to obtain a keyword package comprises the following steps:

[0028] Based on the existing words in the big data, the electronic documents are compared and word segmentation is completed to obtain at least one pending word;

[0029] Acquire at least one daily vocabulary based on big data;

[0030] Counting the frequencies of occurrence of the pending words in the electronic document, sorting and numbering the pending words from the highest to the lowest according to the frequencies of occurrence of the pending words;

[0031] The undetermined words with numbers less than the preset numbers are taken as keywords of the electronic document and summarized to obtain a keyword package.

[0032] Preferably, the matching of at least one keyword to be screened based on the search term and keyword similarity comparison mechanism comprises the following steps:

[0033] Obtain a vocabulary database based on big data, use a keyword similarity comparison mechanism, and calculate the similarity between the search word and the words in the vocabulary database;

[0034] In the vocabulary database, words whose similarity with the search word is greater than the similarity allowable coefficient are obtained as keywords to be screened.

[0035] Preferably, forming at least one search package based on possible combinations of at least one target keyword comprises the following steps:

[0036] Obtaining the number of at least one target keyword as the target number;

[0037] At least one target keyword is randomly combined to obtain at least one search package, wherein the number of the target keywords in the search package ranges from 0 to a target number.

[0038] Preferably, the step of sorting the search packets to obtain a search packet sequence comprises the following steps:

[0039] The search weights of the target keywords in the search package are accumulated and averaged to obtain the search package weight;

[0040] The search packages are sorted from large to small according to their weights to obtain a search package sequence.

[0041] Preferably, obtaining at least one electronic document corresponding to the search package as a target electronic document comprises the following steps:

[0042] Count the occurrence frequency of the target keywords in the search package in the electronic documents as the target frequency;

[0043] Based on the target frequency and the number of bytes of the target keyword, a target total number of bytes in which the target keyword appears in the electronic document is obtained;

[0044] The target total number of bytes is divided by the number of bytes of the electronic document to obtain the target percentage of the target keyword;

[0045] The target proportions of the target keywords in the search package are accumulated and averaged to obtain the average proportion;

[0046] The electronic documents whose average proportion is greater than the preset proportion are taken as target electronic documents, and the target electronic documents are paired with the search package.

[0047] Preferably, the step of sorting the target electronic documents to obtain a target electronic document sequence comprises the following steps:

[0048] Obtaining a search package weight of a search package corresponding to a target electronic document as a target search package weight;

[0049] Multiply the target search package weight by the average proportion of the target electronic documents to obtain the ranking value of the target electronic documents;

[0050] The target electronic documents are arranged from large to small according to the target electronic document ranking values ​​to obtain a target electronic document sequence.

[0051] Preferably, generating the key contents of the common part and the key contents of the different part of the target electronic document comprises the following steps:

[0052] taking a sentence in the target electronic document that contains the target keyword in the search package corresponding to the target electronic document as the target sentence;

[0053] De-duplicate the target sentence to obtain the target de-duplicate sentence;

[0054] Comparing at least one target deduplication sentence of a target electronic document to obtain at least one common target sentence, wherein each of the target electronic documents has a target deduplication sentence whose overlap with the common target sentence is higher than a preset value;

[0055] Summarize at least one common target sentence as key content of the common part of the target electronic document;

[0056] At least one target deduplication sentence in the target electronic document whose overlap degree with at least one common target sentence is less than a preset value is used as the key content of the difference part of the target electronic document.

[0057] An electronic document intelligent screening and query system, used to implement the above-mentioned electronic document intelligent screening and query method, comprises:

[0058] A data forming module, wherein the data forming module aggregates at least one electronic document to generate an electronic document database;

[0059] An interface generation module, wherein the interface generation module generates a query interface;

[0060] A mechanism establishment module, wherein the mechanism establishment module establishes a keyword similarity comparison mechanism, and the keyword similarity comparison mechanism is determined by a similarity allowable coefficient;

[0061] A keyword extraction module, which generates keywords for electronic documents in the electronic document database to obtain keyword packages, and pairs the electronic documents with the keyword packages;

[0062] A known query module, when the queryer knows the name of the electronic document, the known query module obtains the name of the electronic document input by the queryer and searches for the corresponding electronic document;

[0063] an unknown query module, which, when the inquirer does not know the name of the electronic document, obtains at least one search term input by the inquirer, matches at least one to-be-screened keyword based on a similarity comparison mechanism between the search term and the keyword, and obtains at least one target keyword determined by the inquirer from the at least one to-be-screened keyword;

[0064] A weight acquisition module, wherein the weight acquisition module acquires the search weight of the target keyword set by the inquirer;

[0065] A search package forming module, the search package forming module forming at least one search package based on possible combinations of at least one target keyword;

[0066] a sorting module, wherein the sorting module sorts the search packages based on the search weight of the target keyword to obtain a search package sequence, arranges the search packages in order at the top of the display frame according to the search package sequence, obtains at least one electronic document corresponding to the search package as a target electronic document using a keyword similarity comparison mechanism, sorts the target electronic documents based on the search weight of the target keyword to obtain a target electronic document sequence, and displays the target electronic documents in order of the target electronic document sequence at the bottom of the search package display area;

[0067] A difference preview module, wherein the difference preview module generates key contents of a common part and key contents of a difference part of a target electronic document, wherein the key contents of the common part are displayed in a search package display area, and the key contents of the difference part are displayed in a display area corresponding to the target electronic document;

[0068] A feedback adjustment module is used to adjust the similarity allowable coefficient based on the feedback of the query result by the queryer, and to perform a re-query using the adjusted similarity allowable coefficient.

[0069] Compared with the prior art, the present invention has the following beneficial effects:

[0070] By setting up a mechanism establishment module, a weight acquisition module, a search package formation module, a difference preview module and a feedback adjustment module, it is possible to intelligently match the query keywords according to the content input by the inquirer, and form a variety of search packages. The search packages are used to retrieve electronic documents, thereby obtaining a variety of search results so that the inquirer can obtain the electronic documents he needs. According to the weight set by the inquirer, the search content is sorted so that the content with high relevance can be displayed first. In addition, the key content of the common part and the key content of the difference part of the target electronic document are displayed, so that the inquirer can quickly compare and complete the screening of electronic documents. At the same time, when the search displays too many or too few results, the search accuracy can be adjusted autonomously, and then the search can be re-performed to obtain the required electronic documents. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 It is a flow chart of the electronic document intelligent screening and query method of the present invention;

[0072] Figure 2 A schematic diagram of a flow chart of establishing a keyword similarity comparison mechanism of the present invention;

[0073] Figure 3 A schematic diagram of a flow chart of generating keywords for electronic documents in an electronic document database to obtain a keyword package according to the present invention;

[0074] Figure 4 A schematic diagram of a process of matching at least one keyword to be screened based on a search term and keyword similarity comparison mechanism of the present invention;

[0075] Figure 5 A schematic diagram of a process of forming at least one search package based on possible combinations of at least one target keyword according to the present invention;

[0076] Figure 6 A schematic diagram of a process of sorting search packets to obtain a search packet sequence according to the present invention;

[0077] Figure 7A schematic diagram of a process of obtaining at least one electronic document corresponding to a search package as a target electronic document according to the present invention;

[0078] Figure 8 A schematic diagram of a process of sorting target electronic documents to obtain a target electronic document sequence according to the present invention;

[0079] Fig. 9 It is a schematic diagram of a flow chart of generating key contents of a common part and key contents of a different part of a target electronic document according to the present invention. DETAILED DESCRIPTION

[0080] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0081] Reference Figure 1 As shown, an electronic document intelligent screening and query method includes:

[0082] Aggregating at least one electronic document to generate an electronic document database;

[0083] Generate a query interface, the query interface comprising a search box and at least one display box, the at least one display box being arranged in parallel;

[0084] Establish a keyword similarity comparison mechanism, which is determined by the similarity allowable coefficient;

[0085] Generating keywords for electronic documents in the electronic document database to obtain keyword packages, and pairing the electronic documents with the keyword packages;

[0086] When the queryer knows the name of the electronic document, the electronic document name input by the queryer is obtained, and the corresponding electronic document is searched;

[0087] When the inquirer does not know the name of the electronic document, at least one search term input by the inquirer is obtained, and at least one to-be-screened keyword is matched based on a search term and keyword similarity comparison mechanism, and at least one target keyword determined by the inquirer from the at least one to-be-screened keyword is obtained, wherein the to-be-screened keyword and the corresponding search term are similar keywords;

[0088] Obtain the search weight of the target keyword set by the queryer;

[0089] Based on possible combinations of at least one target keyword, forming at least one search package;

[0090] Based on the search weight of the target keyword, the search packages are sorted to obtain a search package sequence, and the search packages are arranged in order at the top of the display box according to the search package sequence;

[0091] Using a keyword similarity comparison mechanism, obtaining at least one electronic document corresponding to the search package as a target electronic document, sorting the target electronic documents based on the search weight of the target keyword to obtain a target electronic document sequence, and displaying the target electronic documents at the bottom of the search package display area in the order of the target electronic document sequence;

[0092] Generate key contents of the common part and key contents of the difference part of the target electronic document, display the key contents of the common part in the search package display area, and display the key contents of the difference part in the display area corresponding to the target electronic document;

[0093] Based on the inquirer's feedback on the query result, the similarity allowance coefficient is adjusted, and the query is re-performed using the adjusted similarity allowance coefficient.

[0094] When the similarity permission coefficient is adjusted, when the similarity permission coefficient is increased, the requirement for similarity increases, and thus the number of electronic documents retrieved and displayed will decrease. This is to reduce the amount of queries by the inquirer when the number of electronic documents displayed by the query is too large. When the similarity permission coefficient is reduced, the requirement for similarity decreases, and thus the number of electronic documents retrieved and displayed will increase. This is to increase the number of electronic documents displayed when the number of electronic documents displayed by the query is too small, so that the inquirer has a greater chance of obtaining the required electronic documents.

[0095] Reference Figure 2 As shown, establishing a keyword similarity comparison mechanism includes the following steps:

[0096] Set the similarity tolerance coefficient for the keyword similarity comparison mechanism;

[0097] Obtain two keywords to be compared, respectively as the first keyword to be compared and the second keyword to be compared;

[0098] Acquire at least one first sample document in which a first keyword to be compared appears;

[0099] Replacing the first keyword to be compared in the first sample document with the second keyword to be compared to obtain a second sample document;

[0100] Retrieve the second sample document based on the big data to determine whether there is an electronic document whose overlap with the second sample document is greater than a preset value. If yes, use the second sample document as the target second sample document; if no, do not do anything.

[0101] Divide the number of target second sample documents by the number of first sample documents to obtain the similarity between the first keyword to be compared and the second keyword to be compared;

[0102] When the similarity between the first keyword to be compared and the second keyword to be compared is greater than the similarity allowable coefficient, the first keyword to be compared and the second keyword to be compared are similar keywords.

[0103] The keyword similarity comparison mechanism determines whether keywords are similar keywords. Because in the query process, due to the unknown nature of electronic documents, it cannot be guaranteed that the keywords used for query are completely consistent with the keywords in the electronic documents. If a completely consistent mechanism is used for judgment, it is very likely that the electronic documents required for query will be missed.

[0104] Reference Figure 3 As shown, generating keywords for electronic documents in an electronic document database to obtain a keyword package includes the following steps:

[0105] Based on the existing words in the big data, the electronic documents are compared and word segmentation is completed to obtain at least one pending word;

[0106] Acquire at least one daily vocabulary based on big data;

[0107] Counting the frequencies of occurrence of the pending words in the electronic document, sorting and numbering the pending words from the highest to the lowest according to the frequencies of occurrence of the pending words;

[0108] The undetermined words with numbers less than the preset numbers are taken as keywords of the electronic document and summarized to obtain a keyword package.

[0109] The generation of keywords is mainly for summarizing electronic documents. Then, during retrieval, keyword packages are used instead of electronic documents, thereby limiting the time spent on retrieval. However, the effect is the same as that of full-text retrieval of electronic documents.

[0110] Reference Figure 4 As shown, based on the search term and keyword similarity comparison mechanism, matching at least one keyword to be screened includes the following steps:

[0111] Obtain a vocabulary database based on big data, use a keyword similarity comparison mechanism, and calculate the similarity between the search word and the words in the vocabulary database;

[0112] In the vocabulary database, words whose similarity with the search word is greater than the similarity allowable coefficient are obtained as keywords to be screened.

[0113] The keyword similarity comparison mechanism is used to obtain the keywords to be screened. When the accuracy of the keyword similarity comparison mechanism is improved, the number of keywords to be screened decreases. When the accuracy of the keyword similarity comparison mechanism is reduced, the number of keywords to be screened increases. In this way, the number of keywords to be screened used for retrieval can be controlled.

[0114] Reference Figure 5As shown, based on the possible combination of at least one target keyword, forming at least one search package includes the following steps:

[0115] Obtaining the number of at least one target keyword as the target number;

[0116] At least one target keyword is randomly combined to obtain at least one search package, wherein the number of the target keywords in the search package ranges from 0 to a target number.

[0117] Different combinations of target keywords may correspond to a search result, and the electronic documents displayed by the search will change. Therefore, they need to be presented for the inquirer to identify. This is more complete than the query results obtained by the inquirer entering keywords independently, and it is more likely to query and obtain the electronic documents they need.

[0118] Reference Figure 6 As shown, sorting the search packets to obtain the search packet sequence includes the following steps:

[0119] The search weights of the target keywords in the search package are accumulated and averaged to obtain the search package weight;

[0120] The search packages are sorted from large to small according to their weights to obtain a search package sequence.

[0121] The purpose of sorting is to display the query results with higher relevance first, so as to facilitate the identification of the queryer.

[0122] Reference Figure 7 As shown, obtaining at least one electronic document corresponding to the search package as a target electronic document includes the following steps:

[0123] Count the occurrence frequency of the target keywords in the search package in the electronic documents as the target frequency;

[0124] Based on the target frequency and the number of bytes of the target keyword, a target total number of bytes in which the target keyword appears in the electronic document is obtained;

[0125] The target total number of bytes is divided by the number of bytes of the electronic document to obtain the target percentage of the target keyword;

[0126] The target proportions of the target keywords in the search package are accumulated and averaged to obtain the average proportion;

[0127] The electronic documents whose average proportion is greater than the preset proportion are taken as target electronic documents, and the target electronic documents are paired with the search package.

[0128] The target electronic document is an electronic document that has a high degree of relevance to the target keyword in the search package, and thus has a high probability of becoming the electronic document required by the queryer.

[0129] Reference Figure 8 As shown, sorting the target electronic documents to obtain a target electronic document sequence includes the following steps:

[0130] Obtaining a search package weight of a search package corresponding to a target electronic document as a target search package weight;

[0131] Multiply the target search package weight by the average proportion of the target electronic documents to obtain the ranking value of the target electronic documents;

[0132] The target electronic documents are arranged from large to small according to the target electronic document ranking values ​​to obtain a target electronic document sequence.

[0133] The purpose of sorting is to display the query results with higher relevance first, so as to facilitate the identification of the queryer.

[0134] Reference Fig. 9 As shown, generating the key contents of the common part and the key contents of the difference part of the target electronic document includes the following steps:

[0135] taking a sentence in the target electronic document that contains the target keyword in the search package corresponding to the target electronic document as the target sentence;

[0136] De-duplicate the target sentence to obtain the target de-duplicate sentence;

[0137] Comparing at least one target deduplication sentence of a target electronic document to obtain at least one common target sentence, wherein each of the target electronic documents has a target deduplication sentence whose overlap with the common target sentence is higher than a preset value;

[0138] Summarize at least one common target sentence as key content of the common part of the target electronic document;

[0139] At least one target deduplication sentence in the target electronic document whose overlap degree with at least one common target sentence is less than a preset value is used as the key content of the difference part of the target electronic document.

[0140] The inquirer can quickly browse the key contents of the common part of the target electronic documents and preliminarily determine the search package he needs. However, the search package has multiple target electronic documents, and these target electronic documents all contain the key contents of the common part, thus forming the key contents of the difference part. The key contents of the difference part of each target electronic document are different. Therefore, the key contents of the difference part can be quickly browsed to determine the target electronic document he needs. In this way, the inquirer does not need to browse the electronic document in its entirety, which can shorten the retrieval time.

[0141] An electronic document intelligent screening and query system, used to implement the above-mentioned electronic document intelligent screening and query method, comprises:

[0142] A data forming module, wherein the data forming module aggregates at least one electronic document to generate an electronic document database;

[0143] An interface generation module, wherein the interface generation module generates a query interface;

[0144] A mechanism establishment module, wherein the mechanism establishment module establishes a keyword similarity comparison mechanism, and the keyword similarity comparison mechanism is determined by a similarity allowable coefficient;

[0145] A keyword extraction module, which generates keywords for electronic documents in the electronic document database to obtain keyword packages, and pairs the electronic documents with the keyword packages;

[0146] A known query module, when the queryer knows the name of the electronic document, the known query module obtains the name of the electronic document input by the queryer and searches for the corresponding electronic document;

[0147] an unknown query module, which, when the inquirer does not know the name of the electronic document, obtains at least one search term input by the inquirer, matches at least one to-be-screened keyword based on a similarity comparison mechanism between the search term and the keyword, and obtains at least one target keyword determined by the inquirer from the at least one to-be-screened keyword;

[0148] A weight acquisition module, wherein the weight acquisition module acquires the search weight of the target keyword set by the inquirer;

[0149] A search package forming module, the search package forming module forming at least one search package based on possible combinations of at least one target keyword;

[0150] a sorting module, wherein the sorting module sorts the search packages based on the search weight of the target keyword to obtain a search package sequence, arranges the search packages in order at the top of the display frame according to the search package sequence, obtains at least one electronic document corresponding to the search package as a target electronic document using a keyword similarity comparison mechanism, sorts the target electronic documents based on the search weight of the target keyword to obtain a target electronic document sequence, and displays the target electronic documents in order of the target electronic document sequence at the bottom of the search package display area;

[0151] A difference preview module, wherein the difference preview module generates key contents of a common part and key contents of a difference part of a target electronic document, wherein the key contents of the common part are displayed in a search package display area, and the key contents of the difference part are displayed in a display area corresponding to the target electronic document;

[0152] A feedback adjustment module is used to adjust the similarity allowable coefficient based on the feedback of the query result by the queryer, and to perform a re-query using the adjusted similarity allowable coefficient.

[0153] Furthermore, the present solution also proposes a storage medium on which a computer-readable program is stored, and when the computer-readable program is called, the above-mentioned electronic document intelligent screening and query method is executed.

[0154] It is understandable that the storage medium may be a magnetic medium, such as a floppy disk, a hard disk, or a magnetic tape; an optical medium, such as a DVD; or a semiconductor medium, such as a solid state drive (SSD).

[0155] In summary, the advantages of the present invention are: by setting a mechanism establishment module, a weight acquisition module, a search package formation module, a difference preview module and a feedback adjustment module, it is possible to intelligently match the query keywords according to the content input by the inquirer, and form a variety of search packages, and use the search packages to retrieve electronic documents, so that a variety of search results can be obtained, so that the inquirer can obtain the electronic documents he needs, and sort the search content according to the weight set by the inquirer, so that the content with high relevance can be displayed first. In addition, the key content of the common part and the key content of the difference part of the target electronic document are displayed, so that the inquirer can quickly compare and complete the screening of the electronic documents. At the same time, when the search display results are too many or too few, the search accuracy can be adjusted autonomously, and then the search can be re-performed to obtain the required electronic documents.

[0156] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. An intelligent screening and query method for electronic documents, characterized in that: include: Aggregating at least one electronic document to generate an electronic document database; Generate a query interface, the query interface comprising a search box and at least one display box, the at least one display box being arranged in parallel; Establish a keyword similarity comparison mechanism, which is determined by the similarity allowable coefficient; Generating keywords for electronic documents in the electronic document database to obtain keyword packages, and pairing the electronic documents with the keyword packages; When the queryer knows the name of the electronic document, the electronic document name input by the queryer is obtained, and the corresponding electronic document is searched; When the inquirer does not know the name of the electronic document, at least one search term input by the inquirer is obtained, and at least one to-be-screened keyword is matched based on a search term and keyword similarity comparison mechanism, and at least one target keyword determined by the inquirer from the at least one to-be-screened keyword is obtained, wherein the to-be-screened keyword and the corresponding search term are similar keywords; Obtain the search weight of the target keyword set by the queryer; Based on possible combinations of at least one target keyword, forming at least one search package; Based on the search weight of the target keyword, the search packages are sorted to obtain a search package sequence, and the search packages are arranged in order at the top of the display box according to the search package sequence; Using a keyword similarity comparison mechanism, obtaining at least one electronic document corresponding to the search package as a target electronic document, sorting the target electronic documents based on the search weight of the target keyword to obtain a target electronic document sequence, and displaying the target electronic documents at the bottom of the search package display area in the order of the target electronic document sequence; Generate key contents of the common part and key contents of the difference part of the target electronic document, display the key contents of the common part in the search package display area, and display the key contents of the difference part in the display area corresponding to the target electronic document; Based on the inquirer's feedback on the query result, the similarity allowance coefficient is adjusted, and the query is re-performed using the adjusted similarity allowance coefficient.

2. The method for intelligent screening and querying of electronic documents according to claim 1, characterized in that: The establishment of a keyword similarity comparison mechanism comprises the following steps: Set the similarity tolerance coefficient for the keyword similarity comparison mechanism; Obtain two keywords to be compared, respectively as the first keyword to be compared and the second keyword to be compared; Acquire at least one first sample document in which a first keyword to be compared appears; Replacing the first keyword to be compared in the first sample document with the second keyword to be compared to obtain a second sample document; Retrieve the second sample document based on the big data to determine whether there is an electronic document whose overlap with the second sample document is greater than a preset value. If yes, use the second sample document as the target second sample document; if no, do not do anything. Divide the number of target second sample documents by the number of first sample documents to obtain the similarity between the first keyword to be compared and the second keyword to be compared; When the similarity between the first keyword to be compared and the second keyword to be compared is greater than the similarity allowable coefficient, the first keyword to be compared and the second keyword to be compared are similar keywords.

3. The method for intelligent screening and querying of electronic documents according to claim 2, characterized in that: Generating keywords for the electronic documents in the electronic document database to obtain a keyword package comprises the following steps: Based on the existing words in the big data, the electronic documents are compared and word segmentation is completed to obtain at least one pending word; Acquire at least one daily vocabulary based on big data; Counting the frequencies of occurrence of the pending words in the electronic document, sorting and numbering the pending words from the highest to the lowest according to the frequencies of occurrence of the pending words; The undetermined words with numbers less than the preset numbers are taken as keywords of the electronic document and summarized to obtain a keyword package.

4. The method for intelligent screening and querying of electronic documents according to claim 3, characterized in that: The matching of at least one keyword to be screened based on the search word and keyword similarity comparison mechanism includes the following steps: Obtain a vocabulary database based on big data, use a keyword similarity comparison mechanism, and calculate the similarity between the search word and the words in the vocabulary database; In the vocabulary database, words whose similarity with the search word is greater than the similarity allowable coefficient are obtained as keywords to be screened.

5. The method for intelligent screening and querying of electronic documents according to claim 4, characterized in that: The forming of at least one search package based on the possible combination of at least one target keyword comprises the following steps: Obtaining the number of at least one target keyword as the target number; At least one target keyword is randomly combined to obtain at least one search package, wherein the number of the target keywords in the search package ranges from 0 to a target number.

6. The method for intelligent screening and querying of electronic documents according to claim 5, characterized in that: The step of sorting the search packets to obtain a search packet sequence comprises the following steps: The search weights of the target keywords in the search package are accumulated and averaged to obtain the search package weight; The search packages are sorted from large to small according to their weights to obtain a search package sequence.

7. The method for intelligent screening and querying of electronic documents according to claim 6, characterized in that: The step of obtaining at least one electronic document corresponding to the search package as a target electronic document comprises the following steps: Count the occurrence frequency of the target keywords in the search package in the electronic documents as the target frequency; Based on the target frequency and the number of bytes of the target keyword, a target total number of bytes in which the target keyword appears in the electronic document is obtained; The target total number of bytes is divided by the number of bytes of the electronic document to obtain the target percentage of the target keyword; The target proportions of the target keywords in the search package are accumulated and averaged to obtain the average proportion; The electronic documents whose average proportion is greater than the preset proportion are taken as target electronic documents, and the target electronic documents are paired with the search package.

8. The method for intelligent screening and querying of electronic documents according to claim 7, characterized in that: The step of sorting the target electronic documents to obtain a target electronic document sequence comprises the following steps: Obtaining a search package weight of a search package corresponding to a target electronic document as a target search package weight; Multiply the target search package weight by the average proportion of the target electronic documents to obtain the ranking value of the target electronic documents; The target electronic documents are arranged from large to small according to the target electronic document ranking values ​​to obtain a target electronic document sequence.

9. The method for intelligent screening and querying of electronic documents according to claim 8, characterized in that: The generating of the key contents of the common part and the key contents of the difference part of the target electronic document comprises the following steps: taking a sentence in the target electronic document that contains the target keyword in the search package corresponding to the target electronic document as the target sentence; De-duplicate the target sentence to obtain the target de-duplicate sentence; Comparing at least one target deduplication sentence of a target electronic document to obtain at least one common target sentence, wherein each of the target electronic documents has a target deduplication sentence whose overlap with the common target sentence is higher than a preset value; Summarize at least one common target sentence as key content of the common part of the target electronic document; At least one target deduplication sentence in the target electronic document whose overlap degree with at least one common target sentence is less than a preset value is used as the key content of the difference part of the target electronic document.

10. An electronic document intelligent screening and query system, used to implement the electronic document intelligent screening and query method according to any one of claims 1 to 9, characterized in that: include: A data forming module, wherein the data forming module aggregates at least one electronic document to generate an electronic document database; An interface generation module, wherein the interface generation module generates a query interface; A mechanism establishment module, wherein the mechanism establishment module establishes a keyword similarity comparison mechanism, and the keyword similarity comparison mechanism is determined by a similarity allowable coefficient; A keyword extraction module, which generates keywords for electronic documents in the electronic document database to obtain keyword packages, and pairs the electronic documents with the keyword packages; A known query module, when the queryer knows the name of the electronic document, the known query module obtains the name of the electronic document input by the queryer and searches for the corresponding electronic document; an unknown query module, wherein when the queryer does not know the name of the electronic document, the unknown query module obtains at least one search term input by the queryer, matches at least one keyword to be screened based on a search term and keyword similarity comparison mechanism, and obtains at least one target keyword determined by the queryer from the at least one keyword to be screened; A weight acquisition module, wherein the weight acquisition module acquires the search weight of the target keyword set by the inquirer; A search package forming module, the search package forming module forming at least one search package based on possible combinations of at least one target keyword; a sorting module, wherein the sorting module sorts the search packages based on the search weight of the target keyword to obtain a search package sequence, arranges the search packages in order at the top of the display frame according to the search package sequence, obtains at least one electronic document corresponding to the search package as a target electronic document using a keyword similarity comparison mechanism, sorts the target electronic documents based on the search weight of the target keyword to obtain a target electronic document sequence, and displays the target electronic documents in order of the target electronic document sequence at the bottom of the search package display area; A difference preview module, wherein the difference preview module generates key contents of a common part and key contents of a difference part of a target electronic document, wherein the key contents of the common part are displayed in a search package display area, and the key contents of the difference part are displayed in a display area corresponding to the target electronic document; A feedback adjustment module is used to adjust the similarity allowable coefficient based on the feedback of the query result by the queryer, and to perform a re-query using the adjusted similarity allowable coefficient.

Images