Semi-automatic search method for global corporate scientific research funding information based on information aggregation

Through information aggregation and data mining technology, keywords in metadata are extracted and aggregated to form an advanced search method, solving the problem of scientific researchers obtaining information funded by global corporate scientific research funds, and achieving efficient, accurate acquisition and dynamic tracking of information.

CN113987325BActive Publication Date: 2024-12-03BEIJING NORMAL UNIV ZHUHAI
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
CN202111290847.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-11-02
Publication Date
2024-12-03
Estimated Expiration
2041-11-02

AI Technical Summary

Technical Problem

It is difficult for scientific researchers to quickly and accurately obtain funding information for global corporate scientific research funds, because the funding information released by enterprises is highly scattered and disordered, and traditional search engine retrieval methods cannot be effectively aggregated and screened.

Method used

The semi-automatic search method based on information aggregation is adopted to extract and aggregate keywords in domestic and foreign metadata to form an advanced search method, and use data mining technology to obtain and filter related information to build a global enterprise scientific research funding information database.

Benefits of technology

It has enabled scientific researchers to efficiently obtain global corporate scientific research fund funding information, and dynamically track and improve scientific research funding information through information aggregation and data mining technology, improving the accuracy and efficiency of information acquisition.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113987325B_ABST
    Figure CN113987325B_ABST
Patent Text Reader

Abstract

The present invention provides a semi-automatic search method for global enterprise scientific research fund information based on information aggregation. The method aggregates enterprise scientific research fund information to extract keywords, effectively improving the problem of highly dispersed and disordered enterprise scientific research fund information, and can efficiently retrieve and obtain effective enterprise scientific research fund information; based on information aggregation, enterprise scientific research fund information is searched and returned through information aggregation technology in three aspects: search term input, data source, and result return, and highly aggregated enterprise scientific research fund information is obtained; the problem of the inclusion boundary of enterprise scientific research fund information is solved, and the inclusion status can be clearly checked according to the crawled results; multiple data sources comprehensively aggregate global enterprise scientific research fund information resources, making the global scientific research fund information more complete.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of information retrieval, and more specifically, to a semi-automatic search method for global enterprise scientific research fund information based on information aggregation. Background Art

[0002] With the continuous development of science and technology and the popularization of the Internet, information is growing at a geometric rate. While the rich information resources can meet our information needs, they also increase the difficulty and time cost of obtaining effective information. Researchers and students have a great demand for research funding. However, the location of research funding published by enterprises on the Internet is not fixed, and the channels and intensity of publicity are relatively limited. The funding notices issued by enterprises cannot be effectively obtained by researchers. How to efficiently obtain information on corporate research funds has become a problem that needs to be solved.

[0003] Keyword search based on mature search engines is clear, flexible and characterized. It is a way to efficiently obtain information from massive websites with personalized query as the core. Its main function is to connect publishing websites with users, and present websites that meet the search requirements through the search engine's own recommendation algorithm, so as to achieve the effect of obtaining target websites from the user's perspective.

[0004] Scientific research grants, especially domestic scientific research grants, are mostly issued by governments and organizations, and their release locations and forms are relatively fixed. The traditional method of collecting such notifications is to set up a web crawler at the location where scientific research grants are frequently released, and to crawl the notification information regularly and then manually screen it.

[0005] The main defects are: (1) The search engine-based method may allow us to obtain global corporate research funding information, but the information is highly scattered and disordered. If researchers do not extract and aggregate corporate research funding information, and aggregate information through multiple channels, it is difficult to quickly and accurately obtain global corporate research funding information. (2) Companies have a higher degree of freedom to publish research funding information than government agencies. Companies have more than one official website, and different companies publish research funding information in different columns. Therefore, the method of obtaining research funding information released by the government and organizations is not effective for companies. Summary of the invention

[0006] The present invention provides a semi-automatic search method for global corporate scientific research fund information based on information aggregation. The method extracts and aggregates information from a large amount of fragmented information and adopts a semi-automatic method to enable scientific researchers to efficiently obtain global corporate scientific research fund information.

[0007] In order to achieve the above technical effects, the technical solution of the present invention is as follows:

[0008] A semi-automatic search method for global corporate scientific research fund information based on information aggregation includes the following steps:

[0009] S1: Aggregate and extract keywords, combine keywords, and stop words from metadata at home and abroad;

[0010] S2: Domestic and foreign keywords, combined keywords, stop words and other restrictions together form a search formula, and multiple search formulas form a search formula queue;

[0011] S3: Aggregation of domestic and foreign data sources, aggregation of multiple data sources; transformation of domestic and foreign search formulas to form more convenient integrated advanced search formulas and URL address search formulas;

[0012] S4: Use domestic and foreign search methods to obtain results, filter out irrelevant information, manually judge and include information on corporate scientific research funds, and extract stop words from the noise information;

[0013] S5: Extract keywords and combine keywords from the identified domestic and foreign corporate scientific research fund funding information.

[0014] Furthermore, in step S1, the specific process of aggregating and extracting keywords, combining keywords, and stop words from domestic metadata is as follows:

[0015] 1) Collect information on scientific research funding from domestic governments and organizations by using third-party information sources;

[0016] 2) Based on the information aggregation method, the keywords of scientific research funding information of domestic governments and organizations are extracted and aggregated, and processed to obtain the initial keywords and keyword group sets:

[0017] Extract the title name set S of scientific research funding information from domestic governments and organizations; aggregate and extract the isomorphic and highly relevant keyword p information from the set S to obtain the keyword set P; calculate the frequency of the aggregated and extracted keywords, and preliminarily assign a weight weight to the keyword p based on the frequency of the keyword p, the weight of the keyword p is p.weight, p.weight=X, (X=1, 2, 3, ..., n);

[0018] 3) When a scientific research grant title has multiple keywords, aggregate and extract the combined keywords to obtain a combined keyword set T; use the statistical method of co-word analysis to count the frequency of two or more words appearing together to calculate the correlation strength between keywords;

[0019] 4) When the strength of association between keywords, the weight of keywords and the frequency of combined keywords are greater than a certain threshold, they will be called in the next search;

[0020] 5) According to the association strength and word frequency of the combined keyword t, the keywords are classified and aggregated into combined keywords.

[0021] Furthermore, in step S1, the specific process of aggregating and extracting keywords, combining keywords, and stop words from foreign metadata is as follows:

[0022] 1) Collect scientific research funding information from foreign companies, governments, and organizations by using third-party information sources;

[0023] 2) Extract and aggregate foreign enterprise scientific research fund information keywords k from the enterprise scientific research fund titles and contents and assign weights;

[0024] 3) Aggregate and extract the keyword information of isomorphic and highly relevant keywords to obtain the keyword set K;

[0025] 4) Record the occurrence frequency of keyword k in the keyword frequency table, and preliminarily assign its weight weight according to the frequency of keyword k, the weight of keyword k.weight, k.weight=X, (X=1, 2, 3, ..., n);

[0026] 5) When a corporate scientific research grant title has multiple keywords, aggregate and extract the combined keywords to obtain a combined keyword set CK; use the statistical method of co-word analysis to count the frequency of two or more words appearing together to calculate the strength of association between keywords;

[0027] 6) When the strength of association between keywords, the weight of keywords and the frequency of combined keywords are greater than a certain threshold, they will be called in the next search;

[0028] 7) Aggregate and analyze foreign scientific research fund information, and classify and aggregate keywords for management.

[0029] Furthermore, in step S2, the specific process of forming the domestic search formula is:

[0030] 1) Using the scientific research funding information of domestic governments and organizations to extract and aggregate, we obtain the initial keyword set P of domestic enterprise scientific research funding funds and the combined keyword T;

[0031] 2) Input and processing of demand information: Manually supplement the demand information, including: company name, domain name, scientific research funding release time information; construct scope restriction words based on demand information;

[0032] 3) Manually create a stop word library. The stop words are collected in two ways: ① manually add stop words, ② extract from the recycling bin under human monitoring;

[0033] 4) Using computer programs, automatically combine and splice keywords, combined keywords, stop words and range-limited words to form a search engine search formula;

[0034] In step S2, the specific process of forming the foreign search formula is:

[0035] 1) Input and process demand information: Manually supplement the demand information, including: company name, domain name, scientific research funding release time information; construct scope restriction words based on demand information;

[0036] 2) Manually create a stop word library: The stop words can be collected in two ways: ① manually add stop words, ② extract from the recycling bin under human supervision;

[0037] 3) Using computer programs, keywords, combined keywords, stop words and range-limiting words are automatically combined and spliced ​​to form a search engine search formula.

[0038] Furthermore, in step S3, the process of aggregating domestic data sources is:

[0039] 1) Data source aggregation: Aggregate and manage multiple data source entrances to achieve dynamic and comprehensive acquisition of corporate scientific research funding information in a continuous and effective manner: search engine entrance aggregation; third-party scientific research fund information release website information source aggregation; foreign corporate scientific research fund information collection website;

[0040] 2) Use search terms to match search result information through search engines: search terms based on advanced search; search terms based on modifying URL parameters;

[0041] 3) Acquire domestic scientific research funding data through data mining technology and advanced information aggregation and processing technology.

[0042] Furthermore, in step S3, the process of aggregating foreign data sources is:

[0043] 1) Data source aggregation: aggregate management of multiple data source entrances to achieve dynamic and comprehensive acquisition of enterprise scientific research information in a continuous and effective manner: search engine entrance aggregation; third-party scientific research fund information release website information source aggregation;

[0044] 2) Use search terms to match search result information through search engines: search terms based on advanced search; search terms based on modifying URL parameters;

[0045] 3) Obtain domestic scientific research funding data through data mining technology and information aggregation processing technology.

[0046] Furthermore, in step S4, the specific process of processing the results obtained by the domestic search formula is:

[0047] 1) Obtain the obtained results, remove duplicates by title or URL, count the number of duplicates, and prioritize the ones with the highest number of duplicates;

[0048] 2) Manually check the results obtained:

[0049] ①: Preliminarily determine from the title whether it is an enterprise scientific research funding fund. If it cannot be determined, proceed to the next step;

[0050] ②: Determine whether it belongs to domestic enterprise scientific research funding based on the content;

[0051] ③: Record after confirmation, aggregate and classify, and include in the enterprise scientific research fund database for further analysis in the future.

[0052] Furthermore, in step S4, the specific process of processing the results obtained by the foreign search formula is:

[0053] 1) Obtain the obtained results, remove duplicates by comparing titles or URLs, count the number of duplicates, and prioritize the ones with the highest number of duplicates;

[0054] 2) Based on the stop word library, filter rules are constructed, and the results or data are presented after being processed by the filter rules;

[0055] 3) Manually check the results obtained:

[0056] ①: Preliminarily determine from the title whether it is an enterprise scientific research funding fund. If it cannot be determined, proceed to the next step;

[0057] ②: Determine from the content whether it is a foreign enterprise scientific research funding fund;

[0058] ③: Record after confirmation, aggregate and classify, and include in the foreign corporate scientific research fund database to facilitate further analysis in the future.

[0059] Furthermore, in step S5, the process of aggregating and extracting information on domestic enterprise scientific research funds is:

[0060] 1) Based on the stop word library, filter rules are constructed, and the results or data collection data are presented after being processed by the filter rules;

[0061] 2) Aggregate and collect from multiple data sources, aggregate and extract keywords p' from the domestic enterprise scientific research funding information obtained after filtering, and combine keywords t';

[0062] 3) Record the frequency of occurrence of keyword p', and preliminarily assign its weight p'.weight based on the frequency n of keyword p', where p'.weight = X, (X = 1, 2, 3, ..., n);

[0063] 4) When a corporate scientific research grant title has multiple keywords, aggregate and extract the combined keywords to obtain a combined keyword set T, and use the statistical method of co-word analysis to count the frequency of two or more words appearing together to calculate the strength of association between keywords;

[0064] 5) When the strength of association between keywords, the weight of keywords and the frequency of combined keywords are greater than a certain threshold, they will be called in the next search;

[0065] 6) Based on the keyword set P of domestic government and organization scientific research funding, delete the keywords unique to the government and organization, add keywords p' extracted from the data of domestic enterprise scientific research funding, and gradually aggregate to form a keyword set P' mainly based on enterprise scientific research funding, which is expressed as follows using the aggregation function:

[0066] Count((unique keywords of PP, p') = P';

[0067] 7) Based on the combination keyword set T of domestic government and organization scientific research funding funds, delete the unique combination keywords of the government and organizations, add the combination keywords p' extracted from the domestic enterprise scientific research funding fund data, and gradually aggregate to form a combination keyword set T' mainly based on enterprise scientific research funding, which is expressed as follows using the aggregation function:

[0068] Count((unique keyword group of TP), t') = T';

[0069] 8) After that, keywords and keyword combinations are extracted and aggregated from each enterprise's scientific research funding information. If new keywords and keyword combinations appear, they are aggregated into the domestic enterprise scientific research funding fund keyword sets P', T' respectively, and the keywords in the field of domestic enterprise scientific research funding funds are continuously improved and refined, so that the keyword information in the field of domestic enterprise scientific research funding funds is highly aggregated.

[0070] Furthermore, in step S5, the process of aggregating and extracting information on foreign enterprise scientific research funds is:

[0071] Aggregate and collect from multiple data sources, aggregate and extract keyword k from the filtered domestic enterprise scientific research funding information, combine keyword ck, and extract stop words from the filtered noise information:

[0072] 1) Aggregation extraction of keyword k:

[0073] ①: Aggregate and analyze the scientific research funding information of domestic enterprises that have been determined to extract keyword k;

[0074] ②: Calculate the frequency of the extracted keyword k and assign weights based on the frequency;

[0075] ③: Compare keyword k with keyword set K to remove duplicates and import into keyword set K;

[0076] 2) Aggregate extraction of combined keyword ck:

[0077] ①: Aggregate and analyze the scientific research funding information of domestic enterprises that have been determined, and extract the combined keyword ck;

[0078] ②: Calculate the frequency of the extracted combined keywords ck and assign weights based on the frequency;

[0079] ③: Compare the combined keyword ck with the keyword set CK to remove duplicates and then import it into the keyword set CK;

[0080] 3) Find stop words in the noise information that has been filtered and determined not to be from corporate scientific research funds;

[0081] 4) After that, extract and aggregate keywords from each company's scientific research funding information. If new keywords appear, aggregate them into the keyword set of foreign company scientific research funding funds, continuously improve and refine the keywords in the field of foreign company scientific research funding funds, so that the keyword information in the field of foreign company scientific research funding funds is highly aggregated;

[0082] 5) Aggregation of Chinese and foreign information, aggregation of domestic and foreign corporate scientific research fund information to build a global corporate scientific research fund information database.

[0083] Compared with the prior art, the technical solution of the present invention has the following beneficial effects:

[0084] The present invention provides a semi-automatic search method for global corporate scientific research fund information based on information aggregation. First, relatively complete scientific research funding information of domestic governments, social organizations, and foreign companies is aggregated and collected as metadata through third-party information sources, and isomorphic and highly matched research funding field theme keywords are extracted from the metadata. Then, the keywords are analyzed, clustered, and aggregated to obtain a search formula suitable for a search engine. Based on the concept of information aggregation, information sources are aggregated, and the aggregation of corporate scientific research fund information is achieved by aggregating multiple data sources with corporate scientific research fund information. Using data mining technology, corporate scientific research funding information is obtained, and aggregation analysis is performed again to collect keywords in the corporate scientific research funding field to construct a unique corporate scientific research funding search. By gradually improving the corporate scientific research fund search formula and continuously iterating, timely scientific research funding information is obtained for scientific researchers, and the dynamic aggregation of scientific research funding information is continuously tracked. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0086] The drawings are for illustrative purposes only and should not be construed as limiting the present patent;

[0087] In order to better illustrate the present embodiment, some parts in the drawings may be omitted, enlarged or reduced, and do not represent the size of the actual product;

[0088] It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0089] The technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0090] like Figure 1 As shown, the present invention provides a semi-automatic search method for global enterprise scientific research fund information based on information aggregation, including the processing process of domestic and foreign information, specifically including the following process:

[0091] domestic:

[0092] (1) Metadata acquisition and processing

[0093] (1-1) Collect information on scientific research funding from domestic governments and organizations by utilizing third-party information sources.

[0094] (1-2) Based on the information aggregation method, the keywords of scientific research funding information of domestic governments and organizations are extracted and aggregated, and processed to obtain the initial keywords and keyword group sets

[0095] (1-2-1) Extract the title name set S of scientific research funding information from domestic governments and organizations.

[0096] (1-2-2) The keyword set P is obtained by aggregating and extracting the isomorphic and highly relevant keyword p information from the set S.

[0097] (1-2-3) Calculate the frequency of the keywords extracted through aggregation, and preliminarily assign a weight weight to the keyword p based on its frequency. The weight of the keyword p is p.weight, where p.weight = X (X = 1, 2, 3, ..., n).

[0098] (1-3) When a scientific research grant title has multiple keywords, aggregate and extract the combined keywords to obtain a combined keyword set T. The statistical method of co-word analysis is used to count the frequency of two or more words appearing together to calculate the strength of association between keywords.

[0099] (1-4) When the strength of association between keywords, the weight of keywords and the frequency of combined keywords are greater than a certain threshold, they will be called in the next search.

[0100] (1-5) According to the association strength and frequency of the combined keyword t, the keywords are classified and aggregated into combined keywords

[0101] For example, scientific research funding information can be aggregated and divided into types such as "topics", "projects", "funds", and guidelines.

[0102] (This example is for reference only and cannot be used as fixed content. There are more than four categories of scientific research funding information.)

[0103] Search formula construction

[0104] (2-1) The scientific research funding information of domestic governments and organizations is extracted and aggregated to obtain the initial keyword set P and combined keyword T of domestic corporate scientific research funding funds.

[0105] (2-2) Input and processing of demand information

[0106] (2-2-1) Manually supplement the required information, such as: company name (full name, abbreviation, other names), domain name (domain name of the company), scientific research funding release time and other information.

[0107] (2-2-2) Construct scope restriction words based on demand information

[0108] (2-3) Manually create a stop word library. The stop words are collected in two ways: ① manually add stop words, ② extract them from the recycling station under human supervision (after comparison with the enterprise scientific research fund data).

[0109] (2-4) Using computer programs, keywords, combination keywords, stop words and range-limiting words are automatically combined and concatenated to form a search engine search formula.

[0110] Aggregate search process

[0111] (3-1) Data source aggregation: Aggregate management of multiple data source entrances to achieve dynamic and comprehensive, continuous and effective acquisition of corporate scientific research funding information.

[0112] (3-1-1) Search engine portal aggregation.

[0113] (3-1-2) Aggregation of information sources from third-party scientific research fund information release websites.

[0114] (3-1-3) Website that collects information on scientific research funds funded by foreign companies.

[0115] (3-2) Use search terms to match result information through search engines.

[0116] (3-2-1) Construction of search formula based on advanced search.

[0117] For example: Search formula: 'site:(xxx.com)title:(company name + scientific research fund "b"(c)-(d))', where 'title:(a"b"(c|d)-(e))' means that the search results contain all the keywords "company name + scientific research fund", include the complete keyword b, include any keyword c, and disable the keyword e. "site:(xxx.com)" means that the website of the specified company organization to be searched is xxx.com.

[0118] (3-2-2) Search formula construction based on modified URL parameters. Construct a new URL search formula by modifying the URL parameter xxx.

[0119] For example:

[0120] URL="https: / / "+xxx+" / s?wd="+xxx+"&pn="+xxx+"&oq="+xxx

[0121] (3-3) Acquire domestic scientific research funding data through advanced technologies such as data mining and information aggregation processing.

[0122] (4) Processing of acquired data.

[0123] (4-1) Obtain the obtained results, remove duplicates by title or URL, count the number of duplicates, and prioritize the ones with the highest number of duplicates.

[0124] (4-2) Manually check the obtained results.

[0125] (4-2-1) Preliminarily determine from the title whether it is a corporate scientific research funding fund. If you cannot determine, proceed to (4-2-2).

[0126] (4-2-2) Determine from the content whether it is a domestic enterprise scientific research funding fund.

[0127] (4-2-3) After confirmation, the records are aggregated and classified and then included in the enterprise scientific research fund database for further analysis in the future.

[0128] (5) Aggregate and extract information from the included corporate scientific research funds.

[0129] (5-1) Based on the stop word library, filter rules are constructed, and the results or data collected after being processed by the filter rules are presented.

[0130] (5-2) Aggregate and collect from multiple data sources, aggregate and extract keyword p' from the domestic enterprise scientific research funding information obtained after filtering, and combine keyword t'.

[0131] (5-3) The frequency of occurrence of keyword p' is recorded, and a weight p'.weight is preliminarily assigned to it based on the word frequency n of keyword p', where p'.weight = X, (X = 1, 2, 3, ..., n).

[0132] (5-4) When a corporate scientific research grant title has multiple keywords, aggregate and extract the combined keywords to obtain a combined keyword set T. The statistical method of co-word analysis is used to count the frequency of two or more words appearing together to calculate the strength of association between keywords.

[0133] (5-5) When the strength of association between keywords, the weight of keywords and the frequency of combined keywords are greater than a certain threshold, they will be called in the next search.

[0134] (5-6) Based on the keyword set P of domestic government and organization scientific research funding, delete the keywords unique to the government and organization, add keywords p' extracted from the data of domestic enterprise scientific research funding, and gradually aggregate to form a keyword set P' mainly based on enterprise scientific research funding. The aggregation function is expressed as follows:

[0135] Count((unique keywords of PP, p') = P'

[0136] (5-7) Based on the combination keyword set T of domestic government and organization scientific research funding funds, delete the combination keywords unique to the government and organizations, add the combination keywords p' extracted from the domestic enterprise scientific research funding fund data, and gradually aggregate to form a combination keyword set T' mainly based on enterprise scientific research funding. The aggregation function is expressed as follows:

[0137] Count((TP's unique keyword group), t') = T'

[0138] (5-8) After that, keywords and keyword combinations are extracted and aggregated from each company's scientific research funding information. If new keywords and keyword combinations appear, they are aggregated into the domestic company scientific research funding fund keyword set P', T' respectively, and the keywords in the field of domestic company scientific research funding funds are continuously improved and refined, so that the keyword information in the field of domestic company scientific research funding funds is highly aggregated.

[0139] foreign:

[0140] (6) Metadata acquisition and processing

[0141] (6-1) Collect scientific research funding information from foreign companies, governments, and organizations by utilizing third-party information sources.

[0142] (6-2) Extract and aggregate foreign corporate scientific research fund information keywords k from the corporate scientific research fund titles and contents

[0143] Assign weight.

[0144] (6-3) Aggregate and extract the homogeneous and highly relevant keyword information to obtain the keyword set K.

[0145] (6-4) The frequency of occurrence of keyword k is recorded in the keyword frequency table, and a weight weight is preliminarily assigned to keyword k based on its word frequency, the weight of the keyword k.weight, k.weight = X, (X = 1, 2, 3, ..., n).

[0146] (6-5) When a corporate scientific research grant title has multiple keywords, aggregate and extract the combined keywords to obtain a combined keyword set CK. The statistical method of co-word analysis is used to count the frequency of two or more words appearing together to calculate the strength of association between keywords.

[0147] (6-6) When the strength of association between keywords, the weight of keywords and the frequency of combined keywords are greater than a certain threshold, they will be called in the next search.

[0148] (6-7) Aggregate and analyze foreign scientific research fund information and classify and aggregate keywords for management.

[0149] For example, scientific research funding information can be aggregated into types such as "grants", "program", "awards", and "fund".

[0150] (This example is for reference only and cannot be used as fixed content. There are more than four categories of scientific research funding information.)

[0151] (7) Construction of search formula

[0152] (7-1) Input and processing of demand information

[0153] (7-1-1) Manually supplement the required information, such as: company name (full name, abbreviation, other names), domain name (domain name of the company), scientific research funding release time and other information.

[0154] (7-1-2) Construct scope restriction words based on demand information.

[0155] (7-2) Manually create a stop word library. The stop words are collected in two ways: ① manually add stop words, ② extract them from the recycling station under human supervision (after comparison with the enterprise scientific research fund data).

[0156] (7-3) Using computer programs, keywords, combination keywords, stop words and range-limiting words are automatically combined and concatenated to form a search engine search formula.

[0157] (8) Aggregate search process

[0158] (8-1) Data source aggregation: Aggregate management of multiple data source entrances to achieve dynamic, comprehensive, continuous and effective acquisition of corporate scientific research information.

[0159] (8-1-1) Search engine portal convergence.

[0160] (8-1-2) Aggregation of information sources from third-party scientific research fund information release websites.

[0161] (8-2) Use search terms to match result information through search engines.

[0162] (8-2-1) Construction of search formula based on advanced search.

[0163] For example: Search formula: all words any word 1 OR any word 1 "exact match word" - no need for words site:.xxx 2001..2021

[0164] (8-2-2) Search formula construction based on modified URL parameters. Construct a new URL search formula by modifying the URL parameter xxx.

[0165] For example:

[0166] URl = "Https: / / " + xxx (important words) + " / search? as_q = " + xxx (exact match) + "&as_oq = " + xxx (excluding the following words) + "&as_nlo = " + xxx (number range)

[0167] (8-3) Acquire domestic scientific research funding data through advanced technologies such as data mining and information aggregation processing.

[0168] (9) Processing of acquired data.

[0169] (9-1) Obtain the obtained results, remove duplicates by comparing the titles or URLs, and count the number of duplicates, and prioritize the ones with the highest number of duplicates.

[0170] (9-2) Based on the stop word library, filter rules are constructed, and the results or data are presented after being processed by the filter rules.

[0171] (9-3) Manually check the obtained results.

[0172] (9-3-1) Preliminarily determine from the title whether it is a corporate scientific research funding fund. If it cannot be determined, proceed to (9-3-2).

[0173] (9-3-2) Determine from the content whether it is a foreign enterprise scientific research funding fund.

[0174] (9-4) After confirmation, the records are aggregated and classified and included in the database of foreign corporate scientific research funds to facilitate further analysis in the future.

[0175] (10) Aggregate and extract information from the included corporate scientific research funds.

[0176] Aggregate and collect from multiple data sources, aggregate and extract keyword k from the filtered domestic enterprise scientific research funding information, combine keyword ck, and extract stop words from the filtered noise information.

[0177] (10-1) Aggregation extraction of keyword k.

[0178] (10-1-1) Aggregate and analyze the identified domestic enterprise scientific research funding information to extract the keyword k.

[0179] (10-1-2) Calculate the frequency of the extracted keyword k and assign weights based on the frequency.

[0180] (10-1-3) Compare keyword k with keyword set K to remove duplicates and import into keyword set K.

[0181] (10-2) Aggregate extraction of combined keyword ck.

[0182] (10-2-1) Aggregate and analyze the scientific research funding information of domestic enterprises that have been determined, and extract the combined keyword ck.

[0183] (10-2-2) Calculate the frequency of the extracted combined keyword ck and assign weights based on the frequency.

[0184] (10-2-3) The combined keyword ck is compared with the keyword set CK to remove duplicates and then imported into the keyword set CK.

[0185] (10-3) Find stop words in the noise information that has been filtered and determined not to be from corporate scientific research funds.

[0186] (11) After that, keywords are extracted and aggregated from each company's scientific research funding information. If new keywords appear, they are aggregated into the keyword set of foreign company scientific research funding funds, and the keywords in the field of foreign company scientific research funding funds are continuously improved and refined, so that the keyword information in the field of foreign company scientific research funding funds is highly aggregated.

[0187] (12) Aggregation of Chinese and foreign information: Aggregation of information on scientific research funds funded by domestic and foreign enterprises to build a global corporate scientific research fund information database.

[0188] Specific examples:

[0189] 1. Examples of Chinese corporate scientific research funds:

[0190] Taking the scientific research fund of a Fortune 500 company as an example, Chinese keywords are extracted from the government and organization, and aggregated into keyword sets: {fund, project, guide, topic, application, joint, joint fund, project guide, research, science, application, open, innovation, open fund, technology, key laboratory}

[0191] Then, through the association strength and word frequency, the combination keyword set is aggregated into {

[0192] Scientific research fund, joint fund, fund application, research fund, open topic, research topic, development project, research project, cooperative project, plan, cooperative exchange project, funded project, scientific project, engineering project, construction project, technical award, science and technology award; bidding announcement, bidding topic, topic bidding, topic guide

[0193] }

[0194] Form the search term (

[0195] site:(tencent.com) A company fund "Fund" (Fund | Project | Application | Guide) - (Baidu Library | Scientific Research Fund)

[0196] site:(tencent.com) a company fund "fund" (fund | bidding | guide | special project)

[0197] site:(tencent.com) A company's scientific research fund "Fund" (Scientific Research Fund | Joint Fund)

[0198] site:(tencent.com) A company fund "Fund" (Fund Application | Research Fund)

[0199] A company fund "topic" (open topic | research topic)

[0200] A company fund "project" (development project, research project, cooperation project)

[0201] A company fund "project" (plan | cooperation and exchange project | funded project | scientific project)

[0202] A company fund "project" (engineering project | construction project)

[0203] A company fund "tender" (tender notice | tender topic | project tender)

[0204] site:(tencent.com) A company fund "Guide" (Project Guide)

[0205] site:(tencent.com) A company joint fund

[0206] site:(tencent.com) A company opens a fund )

[0208] For example: by searching the term "site:(tencent.com) a certain company's scientific research fund "fund" (scientific research fund | joint fund | fund application)"

[0209] Among the search results, there are 9 scientific research fund information of the enterprise. The keywords "fund application" and the combined keywords "fund (topic) (project)--application" and "college cooperation" can be extracted from these 9 title information. Compared with the keyword library and the combined keyword library respectively, the keywords "fund application" and the combined keywords "fund (topic) (project)...application" already exist in the keyword library, so their identification weights are increased accordingly. If "college cooperation" does not exist in the keyword library, it will be entered into the keyword library and the identification weight is 1.

[0210] 2. Explanation of information on scientific research funding from U.S. companies

[0211] Taking the research fund of a Fortune 500 company as an example, the preliminary keyword library includes (Services

[0212] Grants,Invited,Apply,application,Innovative,Training,Improving,Agreement,Studies,Assistance,Assistance,Agreement,Grant,Project,Cooperative,Support,Research,System,Plan,Program,program,fund)

[0213] Then we get some combined keywords by aggregating keywords (

[0214] Research Grants

[0215] Fellowship Program

[0216] Research cooperation

[0217] research program; research program

[0218] grant application

[0219] Funded projects

[0220] Innovation project (program, plan) funding )

[0222] Form the final search term (

[0223] MERCK Fellowship Program-Google Scholarsite:merckgroup.com

[0224] MERCK Research cooperation,site:merckgroup.com

[0225] MERCK research+program-Google Scholar,

[0226] MERCK grant application - Google Scholar,

[0227] MERCK Funded projects - Google Scholar,

[0228] MERCK Innovation project OR program OR plan funding

[0229] For example, the first "2021Research Grants-Merck KGaA" on the first page of the results retrieved by the search term "MERCK Innovation project OR program OR plan funding" is the company's research grant application information page. From the title, we can aggregate and extract the keyword "Research Grants", but the keyword "ResearchGrants" already exists in the original keyword library, so the identification weight of the keyword "Research Grants" in the keyword library is increased by 1. Click the link to enter the details page, and the page aggregates and proposes keywords again, such as: innovativeresearch, research grants program, "innovative research" This keyword is not in the keyword library, so it is included in the keyword library, and the identification weight is initialized to 1; "research grants program" This keyword is not in the keyword library, so it is included in the keyword library, and the identification weight is initialized to 1. )

[0231] The same or similar reference numerals correspond to the same or similar components;

[0232] The positional relationships described in the drawings are only for illustrative purposes and should not be construed as limiting the present patent.

[0233] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the embodiments here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. A semi-automatic search method for global corporate scientific research fund information based on information aggregation, characterized in that: The following steps are involved: S1: Aggregate and extract keywords, combined keywords and stop words from domestic and foreign metadata; The specific process of aggregating and extracting keywords, combining keywords and stop words in domestic metadata is: 1) Collect scientific research funding information from domestic governments and organizations by using third-party information sources; 2) Based on the information aggregation method, the keywords of scientific research funding information of domestic governments and organizations are extracted and aggregated to obtain the initial keywords and keyword group sets: Extract the title name set S of scientific research fund information funded by domestic governments and organizations; aggregate and extract the isomorphic and highly relevant keywords p from the set S to obtain the keyword set P; calculate the frequency of the keywords extracted by aggregation, and preliminarily assign a weight weight to the keyword p based on its frequency, and the weight of the keyword p is recorded as p.weight; 3) When a scientific research grant title has multiple keywords, aggregate and extract the combined keywords to obtain a combined keyword set T; use the statistical method of co-word analysis to count the frequency of two or more words appearing together to calculate the correlation strength between keywords; 4) When the strength of association between keywords, the weight of keywords and the frequency of combined keywords are greater than a certain threshold, they will be called in the next search; 5) According to the association strength and word frequency of the combined keyword t, the keywords are classified and aggregated into combined keywords; The specific process of aggregating and extracting keywords, combining keywords and stop words in foreign metadata is: 1) Collect scientific research funding information from foreign companies, governments and organizations by using third-party information sources; 2) Extract and aggregate foreign enterprise scientific research fund information keywords k from the titles and contents of enterprise scientific research funds and assign weights; 3) Aggregate and extract the keyword information of isomorphic and highly relevant keywords to obtain the keyword set K; 4) Record the occurrence frequency of keyword k in the keyword frequency table, and preliminarily assign a weight weight to keyword k based on its word frequency. The weight of keyword k is recorded as k.weight; 5) When a corporate scientific research grant title has multiple keywords, aggregate and extract the combined keywords to obtain a combined keyword set CK; use the statistical method of co-word analysis to count the frequency of two or more words appearing together to calculate the strength of association between keywords; 6) When the strength of association between keywords, the weight of keywords and the frequency of combined keywords are greater than a certain threshold, they will be called in the next search; 7) Aggregate and analyze foreign scientific research fund information and classify and aggregate keywords for management; S2: Use domestic and foreign keywords, combined keywords, stop words and range-limiting words to form a search formula, and multiple search formulas form a search formula queue; S3: Aggregation of domestic and foreign data sources: Aggregation of multiple data sources; transformation of domestic and foreign search formulas to form advanced search formulas and URL address search formulas; S4: Use domestic and foreign search methods to obtain results, filter out irrelevant information, manually judge and include corporate scientific research fund information, and extract stop words from the noise information; S5: Extract keywords and combine keywords from the identified domestic and foreign corporate scientific research fund funding information; The process of extracting the funding information of scientific research funds of identified domestic enterprises is as follows: 1) Based on the stop word library, build filtering rules, and present the results or collect data after processing by the filtering rules; 2) Aggregate and collect from multiple data sources, aggregate and extract keyword p' and combine keyword t' from the domestic enterprise scientific research funding information obtained after filtering; 3) Record the frequency of occurrence of keyword p', and preliminarily assign its weight p'.weight based on the frequency n of keyword p'; 4) When a corporate scientific research grant title has multiple keywords, aggregate and extract the combined keywords to obtain a combined keyword set T, and use the statistical method of co-word analysis to count the frequency of two or more words appearing together to calculate the strength of association between keywords; 5) When the strength of association between keywords, the weight of keywords and the frequency of combined keywords are greater than a certain threshold, they will be called in the next search; 6) Based on the keyword set P of scientific research funding funds of domestic governments and organizations, delete the unique keywords of governments and organizations, add keywords p' extracted from the data of scientific research funding funds of domestic enterprises, and gradually aggregate to form a keyword set P' mainly based on scientific research funding of enterprises, which is expressed as follows using the aggregation function: Count((unique keywords of PP), p') = P'; 7) Based on the combination keyword set T of scientific research funding funds of domestic governments and organizations, delete the unique combination keywords of governments and organizations, add the combination keywords t' extracted from the aggregation of domestic enterprise scientific research funding funds data, and gradually aggregate to form a combination keyword set T' mainly based on enterprise scientific research funding, which is expressed as follows using the aggregation function: Count((unique keyword group of TP), t') = T'; 8) After that, extract and aggregate keywords and combination keywords from each collected scientific research funding information of an enterprise. If new keywords and combination keywords appear, aggregate them into the domestic enterprise scientific research funding keyword set P' and combination keyword set T' respectively, and continuously improve and refine the keywords in the field of domestic enterprise scientific research funding, so that the keyword information in the field of domestic enterprise scientific research funding is highly aggregated; The process of extracting the funding information of scientific research funds of confirmed foreign enterprises is as follows: Aggregate and collect data from multiple data sources, filter and aggregate the keywords k' and combined keywords ck' in the obtained foreign enterprise scientific research funding information, and extract stop words from the filtered noise information: 1) Aggregation extraction of keyword k': ①: Aggregate and analyze the scientific research funding information of foreign enterprises that have been determined to extract the keyword k'; ②: Calculate the frequency of the extracted keyword k' and assign weights based on the frequency; ③: Compare keyword k' with keyword set K' and remove duplicates, then import into keyword set K'; 2) Aggregate extraction of combined keyword ck': ①: Aggregate and analyze the confirmed foreign enterprise scientific research funding information to extract the combined keyword ck'; ②: Calculate the frequency of the extracted combined keyword ck' and assign weights based on the frequency; ③: Compare the combined keyword ck' with the combined keyword set CK' and remove the duplicates, then import the combined keyword set CK'; 3) Find stop words in the noise information that has been filtered and determined not to be from corporate scientific research funds; 4) After that, extract and aggregate keywords from each collected scientific research funding information of an enterprise. If new keywords appear, aggregate them into the keyword set of foreign enterprise scientific research funding funds, continuously improve and refine the keywords in the field of foreign enterprise scientific research funding funds, so that the keyword information in the field of foreign enterprise scientific research funding funds is highly aggregated; 5) Aggregate information from China and abroad, aggregate domestic and foreign corporate scientific research fund funding information, and build a global corporate scientific research fund information database.

2. The semi-automatic search method for global corporate scientific research fund information based on information aggregation according to claim 1 is characterized in that: In step S2, the specific process of forming the domestic search formula is: 1) Using the scientific research funding information of domestic governments and organizations to extract and aggregate, we obtain the initial keyword set P of domestic enterprise scientific research funding funds and the combined keyword T; 2) Input and processing of demand information: Manually supplement demand information, including: company name, domain name, scientific research funding release time information; construct scope restriction words based on demand information; 3) Manually create a stop word library. The stop words are collected in two ways: ① manually add stop words, ② extract from the recycling bin under human monitoring; 4) Using computer programs, automatically combine and splice keywords, combination keywords, stop words and range-limiting words to form a search engine search formula; In step S2, the specific process of forming the foreign search formula is: 1) Input and processing of demand information: Manually supplement demand information, including: company name, domain name, scientific research funding release time information; construct scope restriction words based on demand information; 2) Manually create a stop word library: The stop words can be collected in two ways: ① manually add stop words, ② extract from the recycling bin under human supervision; 3) Using computer programs, keywords, combination keywords, stop words and range-limiting words are automatically combined and spliced ​​to form a search engine search formula.

3. The semi-automatic search method for global enterprise scientific research fund information based on information aggregation according to claim 2 is characterized in that: In step S3, the process of aggregating domestic data sources is: 1) Data source aggregation: aggregate management of multiple data source entrances: including search engine entrances, third-party scientific research fund information release website information sources, and domestic enterprise scientific research fund funding information collection websites; 2) Use search terms to match search result information through search engines: search terms based on advanced search; search terms based on modifying URL parameters; 3) Acquire domestic scientific research funding data through data mining technology and information aggregation processing advanced technology.

4. The semi-automatic search method for global enterprise scientific research fund information based on information aggregation according to claim 3 is characterized in that: In step S3, the process of aggregating foreign data sources is: 1) Data source aggregation: aggregate management of multiple data source entrances: including search engine entrances and third-party scientific research fund information release website information sources; 2) Use search terms to match search result information through search engines: search terms based on advanced search; search terms based on modifying URL parameters; 3) Acquire foreign scientific research funding data through data mining technology and information aggregation processing technology.

5. The semi-automatic search method for global enterprise scientific research fund information based on information aggregation according to claim 4 is characterized in that: In step S4, the specific process of processing the results obtained by the domestic search formula is: 1) Obtain the obtained results, remove duplicates by title or URL, count the number of duplicates, and prioritize the ones with the highest number of duplicates; 2) Manually check the results obtained: ①: Preliminarily determine from the title whether it is an enterprise scientific research funding fund. If it cannot be determined, proceed to the next step; ②: Determine whether it belongs to domestic enterprise scientific research funding based on the content; ③: Record after confirmation, and include it in the enterprise scientific research fund database after aggregation and classification.

6. The semi-automatic search method for global enterprise scientific research fund information based on information aggregation according to claim 5 is characterized in that: In step S4, the specific process of processing the results obtained by the foreign search formula is: 1) Obtain the obtained results, remove duplicates by comparing titles or URLs, count the number of duplicates, and prioritize the ones with the highest number of duplicates; 2) Based on the stop word library, filter rules are constructed, and the results or data are presented after being processed by the filter rules; 3) Manually check the results obtained: ①: Preliminarily determine from the title whether it is an enterprise scientific research funding fund. If it cannot be determined, proceed to the next step; ②: Determine from the content whether it is a foreign enterprise scientific research funding fund; ③: Record after confirmation, and after aggregation and classification, include it in the foreign enterprise scientific research fund database.

Citation Information

Patent Citations

  • Sci-tech expert information aggregation method based on semantic relation of knowledge organization

    CN106909680A

  • Patent transaction-based retrieval matching method

    CN107492052A

  • Intellectual property big data information retrieval system and storage medium

    CN111737421A

  • Enterprise scientific research achievement management method, an enterprise scientific research achievement management platform, equipment and storage medium

    CN113448918A

  • Method for extracting, merging and ranking search engine results

    US20110040749A1