A management system based on enterprise scientific and technological information achievements

Abstract

The present invention discloses a management system based on enterprise scientific and technological information achievements, including a submission module, an information achievement management system and an information achievement storage database, and the present invention relates to the technical field of scientific and technological achievement management. The management system based on enterprise scientific and technological information achievements is provided with an information achievement management system, and a data acquisition unit is used to classify the collected text feature part once to achieve the matching of the large field features of the prior art, and the classification features are analyzed by the data analysis unit and the data classification unit is used to match the subset of the large field features, so that not only the classification management operation of the enterprise scientific and technological information achievements can be achieved, but also the classification is guaranteed to be more refined, so that personnel can quickly access the achievement files through the existing field feature nouns.

Description

Technical Field

[0001] The present invention relates to the technical field of scientific and technological achievement management, and in particular to a management system based on enterprise scientific and technological information achievements. Background Art

[0002] Referring to the Chinese patent, the patent name is: A management system based on enterprise scientific and technological information achievements (patent publication number: CN112990798A, patent publication date: 2021-06-18), including: data collection module, scientific and technological achievement information database, related achievement analysis module, scientific and technological achievement identification module and related achievement protection module. The data collection module collects the enterprise's scientific and technological achievement autonomous transformation capability data and its corresponding scientific and technological achievement data. After monitoring the theft of the enterprise's scientific and technological achievements, the related achievement analysis module analyzes the degree of correlation between the relevant scientific and technological achievements and the stolen achievements. The related achievement protection module sets different levels of protection measures for the related scientific and technological achievements based on the degree of correlation. The scientific and technological achievement identification module analyzes the technical distance between the enterprise's own capabilities and the value of scientific and technological achievements. According to the size of the technical distance, the appropriate transformation method is selected for the enterprise's various scientific and technological achievements. On the premise of protecting the relevant scientific and technological achievements from being stolen, the security protection cost is saved and the success rate of the transformation of scientific and technological achievements is improved.

[0003] Based on the description in the above documents, when existing enterprises manage the information results of scientific and technological patent innovation, their result files are mixed together after storage, so that when personnel want to review them but do not know the file names, they need to check the files one by one. When the managed result files are classified again, there will be problems with mismatching of information features. For this reason, the present invention provides a management system based on enterprise scientific and technological information results. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides a management system based on enterprise scientific and technological information achievements, which solves the problem that when existing enterprises manage the information achievements of scientific and technological patent innovations, their achievement files are mixed together after storage, so that when personnel want to review them but do not know the file names, they need to check the files one by one, and when the managed achievement files are classified again, there will be a problem of mismatch with information features.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a system based on enterprise scientific and technological information achievement management system, comprising:

[0006] The submission module uploads the scientific and technological patent innovation results files through multiple submission ports and imports them into the enterprise's information results management system;

[0007] The information achievement management system is used to complete the processing of achievement documents and includes:

[0008] The data collection unit is used to collect and classify the text feature parts in the file, and then transmit the classified information to the next unit for analysis;

[0009] The data analysis unit determines whether there is a deviation in the text feature part through the collected text feature part, and extracts the identifiable classification feature and marks it as M n {M1, M2, M3, ..., M n};

[0010] Data classification unit, classifying features M n Compare with the subset features of the large domain features in the prior art and mark the subset features as L n , and finally complete the classification operation of the achievement documents;

[0011] A data storage unit stores, backs up and transmits the achievement files to an information achievement storage database according to the classification results;

[0012] Information results storage database, used for unified storage and backup of processed information data, and for access and sharing of data information;

[0013] The visualization display platform is used to view and display scientific and technological information results, and allows personnel to observe problems in data processing.

[0014] Preferably, the specific operation of collecting and classifying the text feature part in the data collection unit is:

[0015] A1. First, confirm the technical direction of the scientific and technological patent innovation according to the title of the achievement document, and search for the description of the technical field in the achievement document;

[0016] A2. Combine the title of the achievement document with the keywords in the technical field, and introduce the major field features in the prior art and mark them as N. n ;

[0017] A3. By analogy between keywords and large domain features, find domain features that contain keywords or keyword hyper-features, and complete a classification operation on the outcome documents.

[0018] Preferably, the algorithm for determining whether there is a deviation in the text feature part in the data analysis unit is:

[0019] B1. Complete the search of all the text in the result file, and separate the text information in order according to the n-gram model to obtain multiple categories of text features;

[0020] B2. Use the classification model to screen multiple types of text features and remove stop word information in the result document;

[0021] B3. The innovative technical feature words in the search result documents are combined into feature word packages, and the feature word packages are introduced into the separated multiple categories of text features for analogy, and the text feature parts with deviations are corrected.

[0022] Preferably, the processing steps of the n-gram model in B1 are:

[0023] b01. Multiple sliding windows of different lengths are formed according to the byte stream, and the sliding windows are 2-gram, 3-gram, 4-gram and 5-gram respectively;

[0024] b02. Scan the text content step by step in the order from sliding window 5-gram to sliding window 2-gram, and separate the text content according to natural language processing and vocabulary fluency to form text features;

[0025] b03. Finally, we get 2-character text features, 3-character text features, 4-character text features and 5-character text features.

[0026] Preferably, the scanning operation for the text content in b02 is:

[0027] b021. First, slide the sliding window n-gram from the beginning of the text content. If no text features are formed after recognition, the sliding window n-gram moves 1 byte to the right. Otherwise, if text features are formed after recognition, it moves n bytes to the right.

[0028] b022. When recognizing line paragraphs, the sliding window n-gram before reaching the end of the paragraph performs window line division according to the number of words to be divided into lines. The separated window information constitutes a text feature, and the expression of window division is:

[0029] P(n-gram)=(o,no)∑ n [searchn+search*(no)];

[0030] Wherein, P(n-gram) represents the sliding window n-gram, (o, no) represents the segmentation threshold for the sliding window n-gram, o represents the sliding window without line division, and no represents the sliding window after line division, and search n and search(no) represent the word count feature lookup for the sliding window o and the sliding window (no) respectively;

[0031] b023. When the sliding window n-gram moves to the end of a paragraph, if the end of the sliding window n-gram is determined to be a period, the paragraph scanning ends, and all the sliding window n-grams move to the beginning of the next paragraph to start sliding scanning.

[0032] Preferably, the operation of screening multiple types of text features in B2 is:

[0033] c01. First, a classification model for stop words is trained by combining multiple achievement documents, and the commonly used text features in the achievement documents are summarized into the classification model;

[0034] c02. By introducing the obtained text features of various categories into the classification model, the stop words in the text features of various categories are marked;

[0035] c03. Filter the marked stop words and retain the text features of non-stop words.

[0036] Preferably, the step of analogizing the multiple types of text features after the feature vocabulary bag processing in B3 is:

[0037] D1. First, search for specific substitute noun names in the result documents, and combine the substitute noun names into feature vocabulary packages;

[0038] D2. Introduce the feature vocabulary package into the screened multiple categories of text features for comparison, retain the text features that are identical to the noun names in the feature vocabulary package, and review the text features that are partially identical to the noun names in the feature vocabulary package;

[0039] D3. During the review process, the text feature parts that are partially identical to the noun names in the feature vocabulary package and the text feature parts before and after the text feature are checked. If the adjacent text feature parts are combined to obtain a text feature that is identical to the noun name in the feature vocabulary package, the adjacent text feature parts are corrected and combined, and the adjacent parts are separated by the text feature parts and combined into text feature parts respectively, and the final classification feature is marked as M. n {M1, M2, M3, ..., M n}.

[0040] Preferably, the comparison and classification operation in the data classification unit is:

[0041] E1. Summarize and calculate the number of feature words with the same text features in the classification features, and calculate the keywords in the result document based on the frequency of the feature words;

[0042] E2. Combine keywords with the subset features L of the large domain features in the prior art n Make a comparison;

[0043] E3. After comparison, when the keyword and the subset feature L n The same feature or keyword is a subset feature L n When the keyword belongs to the subset feature L n The output document also belongs to the subset feature L n middle.

[0044] Preferably, the frequency value of the feature word is calculated as follows:

[0045]

[0046] Among them, X is the frequency value of the feature word, Y(M n ) represents the number of words in a certain text feature in the classification feature, and Z(M1+M2+…+M n ) is represented as the sum of the number of words in all text features in the classification feature.

[0047] Preferably, the operation of extracting keywords in E1 is:

[0048] F1. After the word frequency of each text feature, compare the word frequency values ​​of each text feature one by one;

[0049] F2, based on the comparison, the three text features with the highest frequency values ​​are used as keywords and used to compare with the existing subset features L n Make a comparison.

[0050] Beneficial Effects

[0051] The present invention provides a system based on enterprise scientific and technological information achievement management system. Compared with the prior art, it has the following beneficial effects:

[0052] (1) The enterprise science and technology information achievement management system is provided with an information achievement management system, and uses a data collection unit to classify the collected text feature part to achieve matching of the existing large-domain features. The classification features are analyzed by the data analysis unit and the data classification unit is used to match the subset of the large-domain features. In this way, not only can the classification management operation of the enterprise science and technology information achievements be realized, but also the classification is guaranteed to be more detailed, so that personnel can quickly review the achievement documents through the existing domain feature terms.

[0053] (2) The enterprise scientific and technological information achievement management system is provided with a data analysis unit, which uses the n-gram model to sequentially separate text information, and uses the classification model to screen multiple categories of text features, removes stop word information in the achievement file, introduces the feature vocabulary package into the separated multiple categories of text features for analogy, and corrects the text feature parts with deviations, so as to achieve more accurate text feature parts during classification, thereby reducing the error rate of text feature separation, so that the achievement files after subsequent classification can correspond to the feature fields.

[0054] (3) The enterprise science and technology information achievement management system is equipped with a data classification unit, which uses the word frequency of feature words to obtain the keywords in the achievement documents, and uses the keywords as classification requirements to compare with the subset features of the large domain features, so as to achieve more efficient and accurate matching of domain features, and realize the rapid classification of achievement documents, thereby reducing the time consumed in subsequent retrieval and improving the efficiency of enterprise science and technology information achievement management. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 It is a principle block diagram of the information achievement management system of the present invention;

[0056] Figure 2 It is a logic judgment diagram of the information achievement management system of the present invention;

[0057] Figure 3 It is a sliding schematic diagram of the sliding window n-gram of the present invention. DETAILED DESCRIPTION

[0058] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0059] See also Figure 1-Figure 3 , the present invention provides two technical solutions:

[0060] Embodiment 1: A management system based on enterprise scientific and technological information achievements, comprising:

[0061] The submission module uploads the scientific and technological patent innovation results files through multiple submission ports and imports them into the enterprise's information results management system;

[0062] The information achievement management system is used to complete the processing of achievement documents and includes:

[0063] The data collection unit is used to collect and classify the text feature parts in the file, and then transmit the classified information to the next unit for analysis;

[0064] The data analysis unit determines whether there is a deviation in the text feature part through the collected text feature part, and extracts the identifiable classification feature and marks it as M n {M1, M2, M3, ..., M n};

[0065] Data classification unit, classifying features M n Compare with the subset features of the large domain features in the prior art and mark the subset features as L n , and finally complete the classification operation of the achievement documents;

[0066] A data storage unit stores, backs up and transmits the achievement files to an information achievement storage database according to the classification results;

[0067] Information results storage database, used for unified storage and backup of processed information data, and for access and sharing of data information;

[0068] The visualization display platform is used to view and display scientific and technological information results, and allows personnel to observe problems in data processing.

[0069] Among them, by setting up an information achievement management system, the data collection unit is used to classify the collected text feature part to achieve the matching of the large-domain features of the existing technology, and then the classification features are analyzed by the data analysis unit and the data classification unit is used to match the subset of the large-domain features. This not only realizes the classification management operation of the enterprise's scientific and technological information achievements, but also ensures that the classification is more detailed, so that personnel can quickly review the achievement documents through the existing domain feature nouns.

[0070] In the embodiment of the present invention, the specific operation of collecting and classifying the text feature part in the data collection unit is as follows:

[0071] A1. First, confirm the technical direction of the scientific and technological patent innovation according to the title of the achievement document, and search for the description of the technical field in the achievement document;

[0072] A2. Combine the title of the achievement document with the keywords in the technical field, and introduce the major field features in the prior art and mark them as N. n ;

[0073] A3. By analogy between keywords and large domain features, find domain features that contain keywords or keyword hyper-features, and complete a classification operation on the outcome documents.

[0074] Among them, a classification operation for the achievement document can be understood in the existing practice: for example, Company A has written a patent achievement document named "A planting device for farmland" through technological innovation, and the characteristic field in the file features of the patent achievement document of "A planting device for farmland" is agricultural planting or planting equipment, and the characteristic field based on the prior art is "agriculture". Therefore, the patent achievement document named "A planting device for farmland" will first be classified into the classification repository with the characteristic field of "agriculture" during classification.

[0075] In the embodiment of the present invention, the algorithm for determining whether there is a deviation in the text feature part in the data analysis unit is:

[0076] B1. Complete the search of all the text in the result file, and separate the text information in order according to the n-gram model to obtain multiple categories of text features;

[0077] B2. Use the classification model to screen multiple types of text features and remove stop word information in the result document;

[0078] B3. The innovative technical feature words in the search result documents are combined into feature word packages, and the feature word packages are introduced into the separated multiple categories of text features for analogy, and the text feature parts with deviations are corrected.

[0079] In the embodiment of the present invention, the processing steps of the n-gram model in B1 are:

[0080] b01. Multiple sliding windows of different lengths are formed according to the byte stream, and the sliding windows are 2-gram, 3-gram, 4-gram and 5-gram respectively;

[0081] b02. Scan the text content step by step in the order from sliding window 5-gram to sliding window 2-gram, and separate the text content according to natural language processing and vocabulary fluency to form text features;

[0082] b03. Finally, we get 2-character text features, 3-character text features, 4-character text features and 5-character text features.

[0083] In the embodiment of the present invention, the scanning operation for the text content in b02 is:

[0084] b021. First, slide the sliding window n-gram from the beginning of the text content. If no text features are formed after recognition, the sliding window n-gram moves 1 byte to the right. Otherwise, if text features are formed after recognition, it moves n bytes to the right.

[0085] b022. When recognizing line paragraphs, the sliding window n-gram before reaching the end of the paragraph performs window line division according to the number of words to be divided into lines. The separated window information constitutes a text feature, and the expression of window division is:

[0086] P(n-gram)=(o,no)∑ n [searchn+search*(no)];

[0087] Wherein, P(n-gram) represents the sliding window n-gram, (o, no) represents the segmentation threshold for the sliding window n-gram, o represents the sliding window without line division, and no represents the sliding window after line division, and search n and search(no) represent the word count feature lookup for the sliding window o and the sliding window (no) respectively;

[0088] b023. When the sliding window n-gram moves to the end of a paragraph, if the end of the sliding window n-gram is determined to be a period, the paragraph scanning ends, and all the sliding window n-grams move to the beginning of the next paragraph to start sliding scanning.

[0089] Among them, by setting up a data analysis unit, the n-gram model is used to separate the text information in order, and the classification model is used to screen the text features of various categories, remove the stop word information in the result file, introduce the feature vocabulary package into the separated multiple categories of text features for analogy, and correct the text feature part with deviations, so as to achieve more accurate text feature part during classification, thereby reducing the error rate of text feature separation, so that the result file after subsequent classification can correspond to the feature field.

[0090] In the embodiment of the present invention, the operation of screening multiple types of text features in B2 is:

[0091] c01. First, a classification model for stop words is trained by combining multiple achievement documents, and the commonly used text features in the achievement documents are summarized into the classification model;

[0092] c02. By introducing the obtained text features of various categories into the classification model, the stop words in the text features of various categories are marked;

[0093] c03. Filter the marked stop words and retain the text features of non-stop words.

[0094] In the embodiment of the present invention, the step of analogizing the multiple types of text features after the feature vocabulary bag is processed in B3 is:

[0095] D1. First, search for specific substitute noun names in the result documents, and combine the substitute noun names into feature vocabulary packages;

[0096] D2. Introduce the feature vocabulary package into the screened multiple categories of text features for comparison, retain the text features that are identical to the noun names in the feature vocabulary package, and review the text features that are partially identical to the noun names in the feature vocabulary package;

[0097] D3. During the review process, the text feature parts that are partially identical to the noun names in the feature vocabulary package and the text feature parts before and after the text feature are checked. If the adjacent text feature parts are combined to obtain a text feature that is identical to the noun name in the feature vocabulary package, the adjacent text feature parts are corrected and combined, and the adjacent parts are separated by the text feature parts and combined into text feature parts respectively, and the final classification feature is marked as M. n {M1, M2, M3, ..., M n}.

[0098] Embodiment 2 is different from Embodiment 1 in that: in the embodiment of the present invention, the comparison and classification operation in the data classification unit is:

[0099] E1. Summarize and calculate the number of feature words with the same text features in the classification features, and calculate the keywords in the result document based on the frequency of the feature words;

[0100] E2. Combine keywords with the subset features L of the large domain features in the prior art n Make a comparison;

[0101] E3. After comparison, when the keyword and the subset feature L n The same feature or keyword is a subset feature L n When the keyword belongs to the subset feature L n The output document also belongs to the subset feature L n middle.

[0102] In the embodiment of the present invention, the calculation formula of the word frequency value of the feature word is:

[0103]

[0104] Among them, X is the frequency value of the feature word, Y(M n ) represents the number of words in a certain text feature in the classification feature, and Z(M1+M2+…+M n ) is represented as the sum of the number of words in all text features in the classification feature.

[0105] In the embodiment of the present invention, the keyword extraction operation in E1 is:

[0106] F1. After the word frequency of each text feature, compare the word frequency values ​​of each text feature one by one;

[0107] F2, based on the comparison, the three text features with the highest frequency values ​​are used as keywords and used to compare with the existing subset features L n Make a comparison.

[0108] Among them, by setting up a data classification unit, the word frequency value of the feature word is used to obtain the keywords in the achievement document, and the keywords are used as classification requirements to compare with the subset features of the large field features, so as to achieve more efficient and accurate matching of field features, and realize the rapid classification of achievement documents, reduce the time consumed in subsequent review, and improve the efficiency of management of enterprise scientific and technological information achievements.

[0109] Based on the keyword extraction of the patent achievement document of "a planting device for farmland" by the above-mentioned company A, the extracted keywords are "planting rack", "feeding pipe" and "feeding frame". By comparing them one by one with the subset features of the characteristic field of "agriculture" in the prior art, the closest subset feature is "feeding, sowing and planting". Therefore, the patent achievement document of "a planting device for farmland" can be divided into the classification repository with the subset feature of "feeding, sowing and planting" again. When checking the achievement document, you only need to search for agriculture, and the category will be displayed. You can find the achievement document by selecting the subset feature title.

[0110] Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0111] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0112] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A system based on enterprise scientific and technological information achievement management system, characterized by: include: The submission module uploads the scientific and technological patent innovation results files through multiple submission ports and imports them into the enterprise's information results management system; The information achievement management system is used to complete the processing of achievement documents and includes: The data collection unit is used to collect and classify the text feature parts in the file, and then transmit the classified information to the next unit for analysis; The data analysis unit determines whether there is a deviation in the text feature part through the collected text feature part, and extracts the identifiable classification feature mark M n ; Data classification unit, classifying features M n Compare with the existing subset features of the large domain features and mark the subset features as L n , and finally complete the classification operation of the achievement documents; A data storage unit stores, backs up and transmits the achievement files to an information achievement storage database according to the classification results; Information results storage database, used for unified storage and backup of processed information data, and for access and sharing of data information; Visual display platform, used to view and display scientific and technological information results, and enable personnel to observe problems in data processing; The algorithm used in the data analysis unit to determine whether there is a deviation in the text feature part is: B1. Complete the search of all the text in the result file, and separate the text information in order according to the n-gram model to obtain multiple categories of text features; B2. Use the classification model to screen multiple types of text features and remove stop word information in the result document; B3. The innovative technical feature words in the search result documents are combined into feature word packages, and the feature word packages are introduced into the separated multiple categories of text features for analogy, and the text feature parts with deviations are corrected; The processing steps of the n-gram model in B1 are: b01. Multiple sliding windows of different lengths are formed according to the byte stream, and the sliding windows are 2-gram, 3-gram, 4-gram and 5-gram respectively; b02. Scan the text content step by step in the order from sliding window 5-gram to sliding window 2-gram, and separate the text content according to natural language processing and vocabulary fluency to form text features; b03, finally obtaining 2-character word features, 3-character word features, 4-character word features and 5-character word features; The scanning operation for the text content in b02 is: b021. First, slide the sliding window n-gram from the beginning of the text content. If no text features are formed after recognition, the sliding window n-gram moves 1 byte to the right. Otherwise, if text features are formed after recognition, it moves n bytes to the right. b022. When recognizing line paragraphs, the sliding window n-gram before reaching the end of the paragraph performs window line division according to the number of words to be divided into lines. The separated window information constitutes a text feature, and the expression of window division is: P(n-gram)=(o,no)Σ n [search o+search(no)]; Wherein, P(n-gram) represents the sliding window n-gram, (o, no) represents the segmentation threshold for the sliding window n-gram, o represents the sliding window without line division, and no represents the sliding window after line division, and search o and search(no) represent the word count feature lookup for the sliding window o and the sliding window (no) respectively; b023. When the sliding window n-gram moves to the end of a paragraph, if the end of the sliding window n-gram is determined to be a period, the paragraph scanning ends, and all the sliding window n-grams move to the beginning of the next paragraph to start sliding scanning.

2. According to claim 1, the enterprise scientific and technological information achievement management system is characterized by: The specific operation of collecting and classifying the text feature part in the data collection unit is as follows: A1. First, confirm the technical direction of the scientific and technological patent innovation according to the title of the achievement document, and search for the description of the technical field in the achievement document; A2. Combine the title of the achievement document with the keywords in the technical field, introduce the existing large field features and mark them as N. n ; A3. By analogy between keywords and large domain features, find domain features that contain keywords or keyword hyper-features, and complete a classification operation on the outcome documents.

3. The enterprise scientific and technological information achievement management system according to claim 1 is characterized by: The operation of screening multiple types of text features in B2 is as follows: c01. First, a classification model for stop words is trained by combining multiple achievement documents, and the commonly used text features in the achievement documents are summarized into the classification model; c02. By introducing the obtained text features of various categories into the classification model, the stop words in the text features of various categories are marked; c03. Filter the marked stop words and retain the text features of non-stop words.

4. According to claim 1, the enterprise scientific and technological information achievement management system is characterized by: The step of analogizing the multiple types of text features processed by the feature vocabulary package in B3 includes: D1. First, search for specific substitute noun names in the result documents, and combine the substitute noun names into feature vocabulary packages; D2. Introduce the feature vocabulary package into the screened multiple categories of text features for comparison, retain the text features that are identical to the noun names in the feature vocabulary package, and review the text features that are partially identical to the noun names in the feature vocabulary package; D3. During the review process, the text feature parts that are partially identical to the noun names in the feature vocabulary package and the text feature parts before and after the text feature will be checked. If the adjacent text feature parts are combined to obtain a text feature that is identical to the noun name in the feature vocabulary package, the adjacent text feature parts will be corrected and combined.

5. The enterprise scientific and technological information achievement management system according to claim 1 is characterized by: The comparison and classification operation in the data classification unit is: E1. Summarize and calculate the number of feature words with the same text features in the classification features, and calculate the keywords in the result document based on the frequency of the feature words; E2. Combine keywords with the existing subset features of large domain features L n Make a comparison; E3. After comparison, when the keyword and the subset feature L n The same feature or keyword is a subset feature L n When the keyword is a subordinate feature of L n In the example, the output document belongs to the subset feature L n middle.

6. The enterprise scientific and technological information achievement management system according to claim 5 is characterized by: The operation of extracting keywords in E1 is: F1. After the word frequency of each text feature, compare the word frequency values ​​of each text feature one by one; F2, based on the comparison, the three text features with the highest frequency values ​​are used as keywords and used to compare with the existing subset features L n Make a comparison.

Images