Document deep retrieval analysis method

By establishing a multi-source database and using keyword matching technology, the problem of low efficiency in traditional document retrieval has been solved, enabling efficient in-depth document retrieval and visual analysis.

CN118939749APending Publication Date: 2024-11-12ZHONGHUIZHI (SHANDONG) HIGH-TECH DEVELOPMENT CO LTD +1

Patent Information

Application Number
CN202410927604.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-11
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

Traditional document retrieval methods are unable to perform in-depth retrieval and analysis, resulting in low retrieval efficiency.

Method used

Establish a multi-source database, filter out the most relevant literature by matching unique identifiers and search keywords, conduct in-depth content analysis, and generate interactive visual analysis reports.

Benefits of technology

It enables in-depth retrieval and analysis of documents, improves retrieval efficiency, and generates visualized analysis reports.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118939749A_ABST
    Figure CN118939749A_ABST
Patent Text Reader

Abstract

The invention discloses a document deep retrieval analysis method, and relates to the technical field of documents. According to the document deep retrieval analysis method, when the documents are retrieved, a database is established through a large number of different documents, a unique identifier is added to each document, then a keyword item is established according to the requirement of a user, three kinds of matching are conducted through the keyword item and the unique identifier, and after matching is completed, the most relevant documents are screened out; compared with the prior art, the method has the advantages that deep retrieval and analysis can be performed on the literature, the efficiency is high, the most relevant literature can be analyzed, the analysis report can be made, and the visual display can be formed. In comparison with the prior art, the method can be used for performing deep retrieval and analysis on the literature, the efficiency is high, and the most relevant literature can be analyzed, the analysis report can be made, and the visual display can be formed.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of document technology, and in particular to a method for in-depth document retrieval and analysis. Background Art

[0002] Documents refer to all carriers that record knowledge. They contain four elements: content with information and knowledge, material carriers that record information and knowledge, recording methods such as text, graphics, audio, and video, and certain forms of expression such as books and journals. Documents can be in various forms such as handwritten, printed, microfilm, audio-visual, or electronic.

[0003] At present, with the continuous expansion of research fields, the amount of literature generated each year is very large. Traditional retrieval methods can usually only perform routine retrieval of literature and cannot perform in-depth retrieval and analysis, resulting in low retrieval efficiency. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present invention provides a method for in-depth retrieval and analysis of documents, which solves the problem that traditional retrieval methods cannot conduct in-depth retrieval and analysis.

[0005] The above technical objectives of the present invention are achieved through the following technical solutions:

[0006] A method for in-depth retrieval and analysis of documents comprises the following steps:

[0007] S1: Collect different literature data and build a multi-source database P based on the literature data y , and for the multi-source database P y Add unique identifiers to different document data in

[0008] S2: Establish search keywords based on user needs x , under the matching algorithm, for the unique identifier and search key words x Match and select the most relevant documents based on the sorting algorithm;

[0009] S3: Conduct in-depth content analysis on the most relevant documents, set a keyword threshold, determine that all terms Aε above the threshold are used as keywords, and combine all keywords into an analysis report;

[0010] S4: The analysis report is converted into interactive charts through Tableau to form a visual display to meet the needs of different users.

[0011] Preferably, the multi-source database P y The steps are as follows:

[0012] S11: Divide different types of document data into multiple categories according to the document names, and mark each category as a1, a2, ..., am, where m is a natural number;

[0013] S12: Assign a unique identifier to each document data under the same category The identifier contains the author name, publication year, document title and keywords of the document data.

[0014] Preferably, the identifier The specific encoding is as follows:

[0015] Identifier Where t is the number of the tth keyword, and t is a natural number;

[0016]

[0017] S121: For each identifier Calculate the hash function, generate a hash value, and add the hash value to each identifier The end of .

[0018] Preferably, the specific calculation method of the hash value is as follows:

[0019] Set the initial hash value to G0 and Iterate, each iteration uses a different hash function Gi to calculate the current identifier The hash value is XORed with the previous hash value;

[0020] After the iteration is completed, the final hash value H∩ is obtained as the identifier The hash value of

[0021] The specific calculation formula is as follows:

[0022]

[0023] in represents XOR operation; Gi is the i-th hash function; is the i-th character in the identifier.

[0024] Preferably, the keyword Ψ x The steps to establish are as follows:

[0025] S21: Based on the text and voice input of the user, convert the voice input into text and mark the initial input text as U y ;

[0026] S22: Initial input text Uy Perform preprocessing, delete pause words and modal particles, and get input text B y ;

[0027] S23: Input text B y Extract keywords and set keyword K as a list, K = {K1, K2, K3, ..., Kn}, sort the keywords in each K and assign a weight value T, T = {T1, T2, T3....., Tn};

[0028] S24: Calculate the score of each keyword based on the frequency and weight of the keyword. The specific calculation is as follows:

[0029] Sc(Kn)=Fr(Kn,By)×Tn, where Sc represents the score, Fr represents the frequency, and Fr(Kn,By) represents the frequency of keyword Kn in the input text B. y The number of times it appears in

[0030] S25: Sort keywords based on the scores of all keywords, and select the ∑th keyword with the highest score, and sort the keyword entry Ψ x Build as follows:

[0031] Ψ x =[Kn|Sc(Kn) is one of the highest ∑ scores]

[0032] S26: multiple search keywords Ψ x Importing multi-source databases y , and in the multi-source database P y Create a table Do in which a unique identifier and a keyword Ψ are included. x fields and create indexes.

[0033] Preferably, the unique identifier and search key words x The specific matching steps are as follows:

[0034] S27: Based on multi-source database P y Establish three types of matches: exact match, partial match, and synonym match;

[0035] The exact match formula is as follows:

[0036]

[0037] M∝ represents the matching result set, which is the collection of all documents that fully match the query conditions;

[0038] d: represents a single document record in a document collection;

[0039] Do represents a table in the database, that is, the document table, which contains all document records;

[0040] A unique identifier for the document Must be equal to the Qi provided by the query;

[0041] Ψ x =Qr_Ψ x Ψ x Must be equal to the Qr_Ψ provided by the query x ;

[0042] ∧: Indicates that the document record will be included in the result set only when both conditions are met at the same time.

[0043] Preferably, the unique identifier and search key words x The partial matching formula is as follows:

[0044]

[0045] Mζ represents the matching result set, i.e., the document set under partial matching conditions;

[0046] d. The keyword entry field indicating the document record contains the keywords provided by the user query;

[0047] The other symbols have the same meaning as for exact match.

[0048] Preferably, the unique identifier and search key words x The synonym matching formula is as follows:

[0049]

[0050] MΩ represents the document set under synonym matching conditions;

[0051] Indicates the existence of synonyms of the query keyword;

[0052] Sy(Qr_Ψ x ) represents a synonym set of the query keyword;

[0053] s∈d.Ψ x Indicates that the keyword entry field of the document record contains a word in the synonym set;

[0054] The other symbols have the same meaning as for safe match and partial match.

[0055] Preferably, the specific steps of the sorting algorithm are as follows:

[0056] S28: For each document, different weights and indicator functions are assigned according to the matching type, and the relevance score is calculated through scoring. The matching document collection is then sorted in descending order according to the score to obtain the most relevant document;

[0057] The formula is as follows:

[0058] Y q (C)=W e ·P iv (d)+W O ·P iv (v)+W b ·P iv (z)

[0059] Y q (C) is the relevance score for a single document C;

[0060] W e , W O , W b The weight coefficients corresponding to complete match, partial match and synonym match respectively, weight the relevance of the document based on the match type, where W e >W O >W b ;

[0061] P iv (d) P iv (v) P iv (z) is an indicator function used to confirm whether document C meets the matching conditions. The specific conditions are as follows:

[0062] P iv (d) indicates that if document C completely matches the query condition, the function value is 1, otherwise it is 0;

[0063] P iv (v) indicates that if document C partially matches the query condition, the function value is 1, otherwise it is 0;

[0064] P iv (z) means that if document C matches the query synonym, the function value is 1, otherwise it is 0.

[0065] The in-depth analysis method of the most relevant literature is as follows:

[0066] S31: Remove non-text elements in the document and segment the text into phrases and words to achieve preprocessing;

[0067] S32: Based on the preprocessed text, extract keywords from the text by using a statistical method;

[0068] The statistical methods are as follows:

[0069] S321: For each document, create a matrix, where the rows represent documents Bα, the columns represent terms Aε (words or phrases), and the cell values ​​represent the frequency of the terms Aε appearing in the document Bα;

[0070] S322: Calculate each term Aε and document Bα, and calculate the frequency IF of the term Aε appearing in document Bα;

[0071] S323: For the term Aε and the document set Dφ, calculate the inverse document frequency IFC;

[0072]

[0073] Where Dφ is the total number of documents in the document set, {Bα∈Dφ:Aε∈Bα} is the document set containing the term Aε;

[0074] S324: Score the importance of each term Aε in each document, and the score is marked as PD;

[0075] PD=IF×IFC

[0076] S325: add up the PD scores and calculate the average value, mark this average value as the threshold, and all scores above this threshold are regarded as keywords;

[0077] S326: Combine all keywords into an analysis report.

[0078] In summary, the present invention mainly has the following beneficial effects:

[0079] When searching for documents, the present application establishes a database through a large number of different documents, adds a unique identifier to each document, and then establishes key words according to user needs. Three types of matching are performed through the key words and the unique identifiers. After the matching is completed, the most relevant documents are screened out, and the most relevant documents are deeply analyzed to prepare an analysis report, and a visual display is formed through interactive charts. Compared with the existing technology, the present application can perform in-depth retrieval and analysis of documents with high efficiency, and can analyze the most relevant documents to prepare an analysis report and form a visual display. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] Figure 1 It is a flow chart of the in-depth retrieval and analysis method of the document of the present invention;

[0081] Figure 2 It is the multi-source database P in the document deep retrieval analysis method of the present invention. y Flow chart of operation steps;

[0082] Figure 3The key words Ψ are established in the literature deep search and analysis method of the present invention. x Flow chart of the steps;

[0083] Figure 4 It is a flowchart of the steps of in-depth analysis of the most relevant documents in the in-depth retrieval and analysis method of the present invention;

[0084] Figure 5 It is a flowchart of the steps of using Tableau in the deep literature retrieval and analysis method of the present invention. DETAILED DESCRIPTION

[0085] The present invention is further described below in conjunction with specific embodiments. However, people familiar with the art should understand that the detailed description given here in conjunction with the drawings is for better explanation, and the structure of the present invention necessarily exceeds these limited embodiments. For some equivalent replacement schemes or common means, they are no longer described in detail herein, but still belong to the scope of protection of the present application.

[0086] Figure 1-5 The best embodiment of the present invention is shown below in conjunction with the attached Figure 1-5 The present invention is further described.

[0087] A method for in-depth retrieval and analysis of documents comprises the following steps:

[0088] S1: Collect different literature data and build a multi-source database P based on the literature data y , and for the multi-source database P y Add unique identifiers to different document data in

[0089] S2: Establish search keywords based on user needs x , under the matching algorithm, for the unique identifier and search key words x Match and select the most relevant documents based on the sorting algorithm;

[0090] S3: Conduct in-depth content analysis on the most relevant documents, set a keyword threshold, determine that all terms Aε above the threshold are used as keywords, and combine all keywords into an analysis report;

[0091] S4: The analysis report is converted into interactive charts through Tableau to form a visual display to meet the needs of different users.

[0092] As a further aspect of the present invention, the multi-source database P y The steps are as follows:

[0093] S11: Divide different types of document data into multiple categories according to the document names, and mark each category as a1, a2, ..., am, where m is a natural number;

[0094] S12: Assign a unique identifier to each document data under the same category The identifier contains the author name, publication year, document title and keywords of the document data.

[0095] As a further feature of the present invention, the identifier The specific encoding is as follows:

[0096] Identifier Where t is the number of the tth keyword, and t is a natural number;

[0097]

[0098] S121: For each identifier Calculate the hash function, generate a hash value, and add the hash value to each identifier The end of .

[0099] As a further aspect of the present invention, the specific calculation method of the hash value is as follows:

[0100] Set the initial hash value to G0 and Iterate, each iteration uses a different hash function Gi to calculate the current identifier The hash value is XORed with the previous hash value;

[0101] After the iteration is completed, the final hash value H∩ is obtained as the identifier The hash value of

[0102] The specific calculation formula is as follows:

[0103]

[0104] in represents XOR operation; Gi is the i-th hash function; is the i-th character in the identifier.

[0105] As a further feature of the present invention, the keyword Ψ x The steps to establish are as follows:

[0106] S21: Based on the text and voice input of the user, convert the voice input into text and mark the initial input text as U y ;

[0107] S22: Initial input text U y Perform preprocessing, delete pause words and modal particles, and get input text B y ;

[0108] S23: Input text B y Extract keywords and set keyword K as a list, K = {K1, K2, K3, ..., Kn}, sort the keywords in each K and assign a weight value T, T = {T1, T2, T3....., Tn};

[0109] S24: Calculate the score of each keyword based on the frequency and weight of the keyword. The specific calculation is as follows:

[0110] Sc(Kn)=Fr(Kn,By)×Tn, where Sc represents the score, Fr represents the frequency, and Fr(Kn,By) represents the frequency of keyword Kn in the input text B. y The number of times it appears in

[0111] S25: Sort keywords based on the scores of all keywords, and select the ∑th keyword with the highest score, and sort the keyword entry Ψ x Build as follows:

[0112] Ψ x =[Kn|Sc(Kn) is one of the highest ∑ scores]

[0113] S26: multiple search keywords Ψ x Importing multi-source databases y , and in the multi-source database P y Create a table Do in which a unique identifier and a keyword Ψ are included. x fields and create indexes.

[0114] As a further feature of the present invention, the unique identifier and search key words x The specific matching steps are as follows:

[0115] S27: Based on multi-source database P y Establish three types of matches: exact match, partial match, and synonym match;

[0116] The exact match formula is as follows:

[0117]

[0118] M∝ represents the matching result set, which is the collection of all documents that fully match the query conditions;

[0119] d: represents a single document record in a document collection;

[0120] Do represents a table in the database, that is, the document table, which contains all document records;

[0121] A unique identifier for the document Must be equal to the Qi provided by the query:

[0122] Ψ x =Qr_Ψ x Ψ x Must be equal to the Qr_Ψ provided by the query x ;

[0123] ∧: Indicates that the document record will be included in the result set only when both conditions are met at the same time.

[0124] As a further feature of the present invention, the unique identifier and search key words x The partial matching formula is as follows:

[0125]

[0126] Mζ represents the matching result set, i.e., the document set under partial matching conditions;

[0127] d. The keyword entry field indicating the document record contains the keywords provided by the user query;

[0128] The other symbols have the same meaning as for exact match.

[0129] As a further feature of the present invention, the unique identifier and search key words x The synonym matching formula is as follows:

[0130]

[0131] MΩ represents the document set under synonym matching conditions;

[0132] Indicates the existence of synonyms of the query keyword;

[0133] Sy(Qr_Ψ x ) represents a synonym set of the query keyword;

[0134] s∈d.Ψ x Indicates that the keyword entry field of the document record contains a word in the synonym set;

[0135] The other symbols have the same meaning as for safe match and partial match.

[0136] As a further aspect of the present invention, the specific steps of the sorting algorithm are as follows:

[0137] S28: For each document, different weights and indicator functions are assigned according to the matching type, and the relevance score is calculated through scoring. The matching document collection is then sorted in descending order according to the score to obtain the most relevant document;

[0138] The formula is as follows:

[0139] Y q (C)=W e ·P iv (d)+W O ·P iv (v)+W b ·P iv (z)

[0140] Y q (C) is the relevance score for a single document C;

[0141] W e , W O , W b The weight coefficients corresponding to complete match, partial match and synonym match respectively, weight the relevance of the document based on the match type, where W e >W O >W b ;

[0142] P iv (d) P iv (v) P iv (z) is an indicator function used to confirm whether document C meets the matching conditions. The specific conditions are as follows:

[0143] P iv (d) indicates that if document C completely matches the query condition, the function value is 1, otherwise it is 0;

[0144] P iv (v) indicates that if document C partially matches the query condition, the function value is 1, otherwise it is 0;

[0145] P iv (z) means that if document C matches the query synonym, the function value is 1, otherwise it is 0.

[0146] As a further aspect of the present invention, the in-depth analysis method of the most relevant documents is as follows:

[0147] S31: Remove non-text elements in the document and segment the text into phrases and words to achieve preprocessing;

[0148] S32: Based on the preprocessed text, extract keywords from the text by using a statistical method;

[0149] The statistical methods are as follows:

[0150] S321: For each document, create a matrix, where the rows represent documents Bα, the columns represent terms Aε (words or phrases), and the cell values ​​represent the frequency of the terms Aε appearing in the document Bα;

[0151] S322: Calculate each term Aε and document Bα, and calculate the frequency IF of the term Aε appearing in document Bα;

[0152] S323: For the term Aε and the document set Dφ, calculate the inverse document frequency IFC;

[0153]

[0154] Where Dφ is the total number of documents in the document set, {Bα∈Dφ:Aε∈Bα} is the document set containing the term Aε;

[0155] S324: Score the importance of each term Aε in each document, and the score is marked as PD;

[0156] PD=IF×IFC

[0157] S325: Add up the PD scores and calculate an average value, mark the average value as a threshold, and all values ​​above the threshold are used as keywords.

[0158] S326: Combine all keywords into an analysis report.

[0159] As a further aspect of the present invention, the steps of using Tableau are as follows:

[0160] S41: Connect Tableau to multiple source databases y , and import the analysis report into Tableau and generate interactive visualization charts;

[0161] S42: Export interactive visualization charts to PDF or images through Tableau to make data easier to transmit.

[0162] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0163] The present invention is described with reference to flowcharts and / or block diagrams of systems, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0164] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0165] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0166] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0167] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for in-depth literature retrieval and analysis. It is characterized in that The following steps are involved: S1: Collect different literature data and build a multi-source database P based on the literature data y , and for the multi-source database P y Add unique identifiers to different document data in S2: Establish search keywords based on user needs x , under the matching algorithm, for the unique identifier and search key words x Match and select the most relevant documents based on the sorting algorithm; S3: Conduct in-depth content analysis on the most relevant documents, set a keyword threshold, determine that all terms Aε above the threshold are used as keywords, and combine all keywords into an analysis report; S4: The analysis report is converted into interactive charts through Tableau to form a visual display to meet the needs of different users.

2. A method for in-depth document retrieval and analysis according to claim 1, It is characterized in that The multi-source database P y The steps are as follows: S11: Divide different types of document data into multiple categories according to the document names, and mark each category as a1, a2, ..., am, where m is a natural number; S12: Assign a unique identifier to each document data under the same category The identifier contains the author name, publication year, document title and keywords of the document data.

3. A method for in-depth document retrieval and analysis according to claim 2, It is characterized in that The identifier The specific encoding is as follows: Identifier = Author's name abbreviation A+publication year P+document mark abbreviation T+keyword N1+keyword N2+…+keyword Nt, where t is the number of the tth keyword, and t is a natural number; S121: For each identifier Calculate the hash function, generate a hash value, and add the hash value to each identifier The end of .

4. A method for in-depth document retrieval and analysis according to claim 3, It is characterized in that The specific calculation method of the hash value is as follows: Set the initial hash value to G0 and Iterate, each iteration uses a different hash function Gi to calculate the current identifier The hash value is XORed with the previous hash value; After the iteration is completed, the final hash value H∩ is obtained as the identifier The hash value of The specific calculation formula is as follows: Where ⊕ represents an XOR operation; Gi is the i-th hash function; is the i-th character in the identifier.

5. A method for in-depth document retrieval and analysis according to claim 1, It is characterized in that The keyword term Ψ x The steps to establish are as follows: S21: Based on the text and voice input of the user, convert the voice input into text and mark the initial input text as U y ; S22: Initial input text U y Perform preprocessing, delete pause words and modal particles, and get input text B y ; S23: Input text B y Extract keywords and set the keyword K as a list, K = {K1, K2, K3, ..., Kn}, sort the keywords in each K and assign a weight value T, T = {T1, T2, T3..., Tn}; S24: Calculate the score of each keyword based on the frequency and weight of the keyword. The specific calculation is as follows: Sc(Kn)=Fr(Kn,By)×Tn, where Sc represents the score, Fr represents the frequency, and Fr(Kn,By) represents the frequency of keyword Kn in the input text B. y The number of times it appears in S25: Sort keywords based on the scores of all keywords, and select the Σth keyword with the highest score, and sort the keyword entry Ψ x Build as follows: Ψ x =[Kn||Sc(Kn) is one of the highest Σ scores] S26: multiple search keywords Ψ x Importing multi-source databases y , and in the multi-source database P y Create a table Do in which a unique identifier and a keyword Ψ are included. x fields and create indexes.

6. A method for in-depth document retrieval and analysis according to claim 1, It is characterized in that The unique identifier and search key words x The specific matching steps are as follows: S27: Based on multi-source database P y Establish three types of matches: exact match, partial match, and synonym match; The exact match formula is as follows: M∝ represents the matching result set, which is the collection of all documents that fully match the query conditions; d: represents a single document record in a document collection; Do represents a table in the database, that is, the document table, which contains all document records; A unique identifier for the document Must be equal to the Qi provided by the query; Ψ x =Qr_Ψ x Ψ x Must be equal to the Qr_Ψ provided by the query x ; ∧: Indicates that the document record will be included in the result set only when both conditions are met at the same time.

7. A method for in-depth document retrieval and analysis according to claim 6, It is characterized in that The unique identifier and search key words x The partial matching formula is as follows: Mζ represents the matching result set, i.e., the document set under partial matching conditions; The keyword entry field indicating the document record contains the keywords provided by the user query; The other symbols have the same meaning as for exact match.

8. A method for in-depth document retrieval and analysis according to claim 6, It is characterized in that The unique identifier and search key words x The synonym matching formula is as follows: MΩ represents the document set under synonym matching conditions; Indicates the existence of synonyms of the query keyword; Sy(Qr_Ψ x ) represents a synonym set of the query keyword; s∈d.Ψ x Indicates that the keyword entry field of the document record contains a word in the synonym set; The other symbols have the same meaning as for safe match and partial match.

9. A method for in-depth document retrieval and analysis according to claim 1, It is characterized in that The specific steps of the sorting algorithm are as follows: S28: For each document, different weights and indicator functions are assigned according to the matching type, and the relevance score is calculated through scoring. The matching document collection is then sorted in descending order according to the score to obtain the most relevant document; The formula is as follows: Y q (C)=W e ·P iv (d)+W O ·P iv (v)+W b ·P iv (z) Y q (C) is the relevance score for a single document C; W e , W O , W b The weight coefficients corresponding to complete match, partial match and synonym match respectively, weight the relevance of the document based on the match type, where W e >W O >W b ; P iv (d) P iv (v) P iv (z) is an indicator function used to confirm whether document C meets the matching conditions. The specific conditions are as follows: P iv (d) indicates that if document C completely matches the query condition, the function value is 1, otherwise it is 0; P iv (v) indicates that if document C partially matches the query condition, the function value is 1, otherwise it is 0; P iv (z) means that if document C matches the query synonym, the function value is 1, otherwise it is 0.

10. A method for in-depth document retrieval and analysis according to claim 1, It is characterized in that The in-depth analysis method of the most relevant literature is as follows: S31: Remove non-text elements in the document and segment the text into phrases and words to achieve preprocessing; S32: Based on the preprocessed text, extract keywords from the text by using a statistical method; The statistical methods are as follows: S321: For each document, create a matrix, in which the rows represent documents Bα, the columns represent terms Aε, and the cell values ​​represent the frequency of the term Aε appearing in the document Bα; S322: Calculate each term Aε and document Bα, and calculate the frequency IF of the term As appearing in document Bα; S323: For the term Aε and the document set Dφ, calculate the inverse document frequency IFC; Where Dφ is the total number of documents in the document set, {Bα∈Dφ:Aε∈Bα} is the document set containing the term Aε; S324: Score the importance of each term Aε in each document, and the score is marked as PD; PD=IF×IFC S325: add up the PD scores and calculate the average value, mark this average value as the threshold, and all scores above this threshold are regarded as keywords; S326: Combine all keywords into an analysis report.

Citation Information

Patent Citations

  • Document similarity calculating method and similar document whole-network retrieval tracking method

    CN106095737A

  • Connection keyword retrieval method based on attribute encryption in hybrid cloud

    CN107948146A

  • Theme and semantic meaning-based dialogue corpus keyword extraction method

    CN108595425A

  • Paper text similarity detection method based on citation network

    CN110489745A

  • Mail contact person recommendation method and device, electronic equipment and storage medium

    CN111934930A

Cited By

  • Digital dynamic map construction method and system for describing literature influence diffusion

    CN119128130A