Auxiliary document reading system based on AI and personalized knowledge base

Abstract

The invention discloses a document auxiliary reading system based on AI and a personalized knowledge base, relates to the technical field of document auxiliary reading, and solves the technical problem that in the prior art, personalized analysis is not carried out on a user, so that the system provides unnecessary auxiliary information, and the reading experience of the user is affected. According to the method, the reading document is converted into the reading text, the historical record of the user is converted into the historical text, the search vocabularies in the historical search text in the historical text are extracted, the historical search vocabularies are obtained in combination with the historical browsing text, the occurrence frequency and the search frequency of the historical search vocabularies are counted, and the historical search record is obtained; and extracting interpretation vocabularies in the reading text, matching the interpretation vocabularies with historical search records, obtaining mastering states of the interpretation vocabularies according to occurrence times and search times of the interpretation vocabularies in the historical search records, generating auxiliary tags based on the mastering states of the interpretation vocabularies, and adding the auxiliary tags to key vocabularies to complete auxiliary reading.

Description

Technical Field

[0001] The present invention belongs to the field of document assisted reading, and in particular is a document assisted reading system based on AI and a personalized knowledge base. Background Art

[0002] When reading text content in a more professional or unfamiliar field, one often needs to understand the relevant background knowledge, especially the meaning of professional terms, so that one can read more smoothly and understand the content; when reading traditional paper text content and books, when encountering unfamiliar professional terms, people often need to use various methods to query relevant knowledge; when reading electronic content online, one can use a computer to help obtain relevant knowledge to assist reading in real time.

[0003] The invention patent with application number CN2023102436014 discloses a method for intelligent assisted reading of professional documents based on a domain knowledge base. The invention constructs a domain ontology knowledge base, builds an intelligent reading assistance platform based on the domain ontology knowledge base, identifies domain concepts through natural language processing technology, and obtains a relational database of the data source, extracts related knowledge in the relational database and combines it with the intelligent reading assistance platform to generate semantic related knowledge, and annotates the document in real time and fills in relevant background knowledge based on the semantic related knowledge; when this method performs assisted reading, since only the relational data of the current document is analyzed, when the user encounters some difficult-to-understand vocabulary or knowledge when reading, in the past reading, by searching for relevant knowledge by themselves, the user may have mastered the corresponding information, and in subsequent reading, there is no need to re-system assisted reading, which causes the system to frequently provide unnecessary auxiliary information, reduce user reading interference, and affect the user's reading experience.

[0004] The present invention provides a document auxiliary reading system based on AI and a personalized knowledge base to solve the above technical problems. Summary of the invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a document assisted reading system based on AI and a personalized knowledge base, which is used to solve the technical problem in the prior art that there is no personalized analysis of users, resulting in the system providing unnecessary auxiliary information, affecting the user's reading experience.

[0006] To achieve the above-mentioned object, the first aspect of the present invention provides a document-assisted reading system based on AI and a personalized knowledge base, comprising: a database, a document conversion module, a document parsing module and a reading-assisted module;

[0007] Document conversion module: used to convert received reading documents into reading text; collect user's historical records and convert them into historical text;

[0008] Document parsing module: used to parse historical texts and extract historical search records; extract interpretation words in reading texts, and obtain the user's mastery of interpretation words based on historical search records and interpretation words;

[0009] Reading assistance module: used to generate auxiliary tags corresponding to interpretation vocabulary based on the mastery status and add them to key vocabulary.

[0010] Preferably, the parsing of historical texts and extracting historical search records includes:

[0011] Extracting historical texts; wherein the historical texts include historical search texts and historical browsing texts;

[0012] The search terms in the historical search text are extracted through natural language processing technology, and the search times of the search terms are recorded; the search terms are matched with the historical browsing text to obtain the historical search terms, and the number of occurrences of the search terms in the historical browsing text is counted, and the number of occurrences and search times of the historical search terms in the historical search text are integrated to obtain the historical search record.

[0013] Preferably, matching the search terms with the historical browsing text to obtain the historical search terms includes:

[0014] A1: Extract search terms and historical browsing text;

[0015] A2: Determine whether the search term appears in the historical browsing text; if yes, mark the corresponding search term as a historical search term; if no, jump to A3;

[0016] A3: Determine whether the search term has the same semantics as the term in the historical browsing text; if yes, mark the corresponding search term as a historical search term; if no, mark the corresponding search term as an irrelevant term.

[0017] Preferably, the step of obtaining the user's mastery of the interpreted vocabulary based on the historical search records and the interpreted vocabulary includes:

[0018] S1: Extract historical search records and interpret vocabulary;

[0019] S2: Determine whether the historical search words in the historical search records contain the interpretation words; if yes, jump to S3; if no, mark the corresponding interpretation words as new words;

[0020] S3: Obtaining the mastery status of the interpreted vocabulary according to the number of appearances and the number of searches of the historical search vocabulary corresponding to the interpreted vocabulary.

[0021] Preferably, the method of obtaining the mastery status of the interpreted vocabulary according to the number of occurrences and the number of searches of the historical search vocabulary corresponding to the interpreted vocabulary includes:

[0022] B1: Extract the occurrence and search times of historical search terms;

[0023] B2: Determine whether the number of searches is greater than 1; if yes, jump to B3; if no, jump to B4;

[0024] B3: Determine whether the number of occurrences is greater than the number of searches; if yes, mark the mastery status of the interpretation vocabulary as mastered; if no, mark the mastery status of the corresponding interpretation vocabulary as difficult to master;

[0025] B4: Determine whether the number of occurrences is greater than the number of searches; if yes, mark the mastery status of the corresponding interpretation vocabulary as fully mastered; if no, mark the mastery status of the corresponding interpretation vocabulary as mastered;

[0026] B5: For the interpretation vocabulary whose mastery status is mastered, set the mastery coefficient, and multiply the ratio of the number of appearances to the number of searches by the mastery coefficient to obtain the mastery value of the interpretation vocabulary; wherein the mastery coefficient is set according to the user's ability to understand the vocabulary;

[0027] B6: Set a mastery threshold; determine whether the mastery value of the interpreted vocabulary is greater than or equal to the mastery threshold; if yes, mark the mastery status of the corresponding interpreted vocabulary as specific mastery; if no, mark the mastery status of the corresponding interpreted vocabulary as rough mastery; wherein the mastery threshold is set based on the average understanding experience of the expert for the user.

[0028] Preferably, the mastery coefficient is set according to the user's ability to understand vocabulary, including:

[0029] Count the historical search words whose appearance times are greater than the search times and mark them as historical mastered words, extract the appearance time and search time of the historical mastered words, and mark the appearance time closest to the current time as the end appearance time, and mark the search time closest to the current time as the end search time;

[0030] Determine whether the terminal appearance time is greater than the terminal search time; if yes, mark the corresponding historically mastered vocabulary as mastered vocabulary; if no, mark the corresponding historically mastered vocabulary as unmastered vocabulary;

[0031] The ratio of the mastered vocabulary to the historically mastered vocabulary is marked as the mastery coefficient.

[0032] Preferably, the generating of auxiliary tags corresponding to the interpretation words based on the mastery status and adding them to the key words includes:

[0033] Extract mastery status;

[0034] The interpretation vocabulary that is difficult to master or roughly mastered is marked as auxiliary reading vocabulary, the text type of the reading text is obtained through natural language processing technology, the professional database corresponding to the text type is matched in the database, the auxiliary reading vocabulary is input into the professional database to obtain relevant information corresponding to the auxiliary reading vocabulary, the relevant information is input into the label file to obtain auxiliary tags, and the auxiliary tags are added to the key vocabulary.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] 1. The present invention converts reading documents into reading texts and converts user historical records into historical texts, extracts search terms in historical search texts in historical texts, combines historical browsing texts to obtain historical search terms, and counts the number of occurrences and searches of historical search terms to obtain historical search records, so as to quickly review important terms and contents searched in the past, save time, and improve work efficiency; extracts interpretation terms in reading texts and matches them with historical search records, obtains the mastery status of interpretation terms according to the number of occurrences and searches of interpretation terms in historical search records, generates auxiliary tags based on the mastery status of interpretation terms and adds them to key terms to complete auxiliary reading, and can provide personalized auxiliary information according to the user's mastery status, provide users with services and contents that better meet their needs, and enhance reading experience.

[0037] 2. The present invention obtains the mastery status of the corresponding interpretation vocabulary by comparing the number of occurrences of historical search words with the number of searches, and analyzes the interpretation words whose mastery status is mastered, compares the terminal appearance time and the terminal search time of the interpretation words, calculates the mastery coefficient, and obtains the mastery value of the interpretation words by combining the mastery coefficient, the number of occurrences and the number of searches. The mastery value is compared with a preset mastery threshold to obtain the mastery status of the interpretation words, and personalized auxiliary information is provided according to the user's actual mastery, so that the user can have a better experience in the reading process. For the words that have been mastered, the system will not frequently provide unnecessary auxiliary information, thereby reducing the user's reading interference and improving reading efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0039] Figure 1 The following is a flowchart of an embodiment of the present invention.

[0040] Figure 2It is a system composition diagram of the present invention.

[0041] Figure 3 The following is a complete flowchart of obtaining the mastering state in one embodiment of the present invention. DETAILED DESCRIPTION

[0042] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than 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.

[0043] See also Figure 1-Figure 3 , the first aspect of the present invention provides a document-assisted reading system based on AI and personalized knowledge base, including: a database, a document conversion module, a document parsing module and a reading assistance module;

[0044] Document conversion module: used to convert the received reading documents into reading text; collect the user's historical records and convert them into historical text.

[0045] For example, in this embodiment, user A wants to read an instruction manual. After the system receives the instruction manual file, it converts the format of the instruction manual file through the text extraction library and extracts the text part therein; if the instruction manual file is a PDF file, the PDF file is converted into a WORD file, and then the text is extracted through the python-docx library to obtain the reading text; the historical records of the user history are extracted and converted into historical text.

[0046] Document parsing module: used to parse historical texts and extract historical search records; extract interpretation words in reading texts, and obtain the user's mastery of interpretation words based on historical search records and interpretation words.

[0047] Exemplarily, natural language processing technology (such as NLP technology) is used to extract words from historical search texts in historical texts and mark them as search words. The number of times the search words appear in the historical search texts is recorded and marked as the number of searches. It is determined whether the search words appear in the historical browsing texts. If the search words appear in the historical browsing texts, it means that the user did not understand the search words when browsing the historical browsing texts and searched for them, and the search words are marked as historical search words. If the search words do not appear in the historical browsing texts, it is determined whether the search words have the same semantics as the words in the historical browsing texts. If there are no words in the historical browsing texts that have the same semantics as the search words, it means that the user has not searched for the search words, and the corresponding search words are marked as irrelevant words. If the words in the historical browsing texts have the same semantics as the search words, it means that the user may have questions about the search words and correct the words and search based on his or her own understanding. For example, "words" and "vocabulary" have the same semantics but different texts. This can prevent the unsuccessful detection of search words due to the user's modification of the original text, improve the accuracy of text recognition, and improve the user's reading experience.

[0048] Through natural language processing technology, the interpretation words in the reading text are refined, and it is determined whether the interpretation words are included in the historical search words. If the interpretation words are not included in the historical search words, it means that the user has not searched for the interpretation words, and the interpretation words are marked as new words; if so, it means that the user has doubts about the interpretation words and searched for them, and it is determined whether the historical search words corresponding to the interpretation words in the historical search records are greater than 1. Specifically, it can be divided into multiple searches and single searches. If it is multiple searches, it is further determined whether the number of occurrences is greater than the number of searches. If the number of occurrences is greater than the number of searches, it means that the user has already understood the word when encountering it later, and the mastery status of the interpretation words is marked as mastered. If the number of occurrences is less than the number of searches, it means that the user has searched for the word multiple times when reading it once, and the mastery of the interpretation words is marked as mastered. The mastery status of the corresponding interpretation vocabulary is marked as difficult to master; if it is a single search, it is determined whether the number of occurrences is greater than the number of searches. If the number of occurrences is greater than the number of searches, it means that the user only searches after the first encounter of the interpretation vocabulary, and does not search for it when it is encountered subsequently. The user has fully mastered the interpretation vocabulary, and the mastery status of the corresponding interpretation vocabulary is marked as fully mastered; if the number of occurrences is not greater than the number of searches, because it is a single search, the number of occurrences and the number of searches are both 1, which means that the user has only encountered the interpretation vocabulary once and searched for it, and the mastery status of the corresponding interpretation vocabulary is marked as mastered; it can accurately judge the user's mastery of the vocabulary based on the user's historical search records and the number of occurrences, provide more accurate auxiliary information, and provide personalized auxiliary information based on the user's actual mastery, so that the user can get a better reading experience during the reading process.

[0049] In some other preferred embodiments, new words may be words that the user has browsed but not searched for. The new words can be matched in the historical browsing text. If the new words appear in the historical browsing text, it means that the user has no questions about the words, and the new words are marked as encountered words; if the new words do not appear in the historical browsing text, it means that the new words are encountered by the user for the first time, and the new words are marked as newly encountered words; in subsequent processing, the encountered words are eliminated, and only the newly encountered words are used for auxiliary reading.

[0050] The historically-mastered vocabulary whose number of appearances is greater than the number of searches is statistically marked as historically-mastered vocabulary, and the appearance time and search time of the historically-mastered vocabulary are extracted, and the appearance time closest to the current time is marked as the end appearance time, and the search time closest to the current time is marked as the end search time; it is determined whether the end appearance time is greater than the end search time. If the end appearance time is greater than the end search time, it means that the historically-mastered vocabulary still appears after the last search, but the user has not searched for it, which means that the historically-mastered vocabulary has been preliminarily mastered, and the historically-mastered vocabulary is marked as a mastered vocabulary; if the end appearance time is not greater than the end search time, it means that the user searched for the historically-mastered vocabulary when reading it, and still searched for it when reading it for the last time, and the historically-mastered vocabulary is marked as an unmastered vocabulary, and the proportion of mastered vocabulary to historically-mastered vocabulary is marked as the mastery coefficient.

[0051] The mastery value of the interpretation vocabulary is obtained by multiplying the ratio of the number of occurrences of the interpretation vocabulary with a mastery status of mastered to the number of searches by the mastery number. In this embodiment, a mastery threshold is set, and the mastery value of the interpretation vocabulary is compared with the mastery threshold. If the mastery value is greater than the mastery threshold, it means that the user has a certain mastery of the interpretation vocabulary and no reading assistance is needed. The mastery status of the corresponding interpretation vocabulary is marked as specific mastery; if the mastery value is less than the mastery threshold, it means that the user's understanding of the interpretation vocabulary is still rough and needs to be further strengthened. The mastery status of the corresponding interpretation vocabulary is marked as rough mastery. Setting the mastery threshold can provide a clear standard for the system to help the system judge the user's mastery of the vocabulary. The user can obtain reading assistance that better meets their needs and improve reading experience and satisfaction. For vocabulary with higher mastery values, the system can reduce or not provide auxiliary information to reduce user reading interference; for vocabulary with lower mastery values, the system can provide more auxiliary information to help users better understand.

[0052] Reading assistance module: used to generate relevant information on vocabulary interpretation based on the mastery status and add it as auxiliary tags to key vocabulary.

[0053] Exemplarily, the interpretation vocabulary that is difficult to master and roughly mastered is marked as auxiliary reading vocabulary, and the text type of the reading text is analyzed through natural language processing technology, and the professional database corresponding to the text type is matched in the professional knowledge database, such as "cooking", "crafts", "electronics" and other professional databases; the auxiliary reading vocabulary is entered into the professional database to obtain the professional information corresponding to the auxiliary reading vocabulary, and the corresponding information is entered into the label file and added to the key vocabulary to complete the auxiliary reading of the document. When the user reads to the key vocabulary, the user can click the label file to read the corresponding professional information to enhance understanding, help users better understand the document and improve the user's reading experience.

[0054] Part of the data in the above formula is calculated by removing the dimension and taking its numerical value. The formula is a formula closest to the actual situation obtained by software simulation of a large amount of collected data; the preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained through simulation of a large amount of data.

[0055] Working principle of the present invention:

[0056] The present invention converts a reading document into a reading text, and converts a user's historical record into a historical text, extracts search terms in the historical search text and records the number of searches, obtains historical search terms by combining historical browsing texts, integrates the number of occurrences of historical search terms with the number of searches to obtain historical search records, extracts interpretation terms in the reading text, matches the interpretation terms with the historical search records, obtains a mastery status based on the number of occurrences and the number of searches of the interpretation terms corresponding to the historical search terms, obtains mastered terms based on the appearance time and search time of the historical search terms, and obtains a mastery coefficient, obtains a mastery status by combining the mastery coefficient, the number of occurrences and the number of searches, generates corresponding data of the interpretation terms in a database, and adds them to the end terms as auxiliary tags to complete auxiliary reading.

[0057] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A document-assisted reading system based on AI and personalized knowledge base, characterized in that: include: Database, document conversion module, document parsing module and reading assistance module; Document conversion module: used to convert the received reading document into reading text; Collect user's history and convert it into historical text; Document parsing module: used to parse historical texts and extract historical search records; extract interpretation words in reading texts, and obtain the user's mastery of interpretation words based on historical search records and interpretation words; Reading assistance module: used to generate auxiliary tags corresponding to interpretation vocabulary based on the mastery status and add them to key vocabulary.

2. The document-assisted reading system based on AI and personalized knowledge base according to claim 1, characterized in that: The historical text is parsed and refined to obtain the historical search records, including: Extracting historical texts; wherein the historical texts include historical search texts and historical browsing texts; The search terms in the historical search text are extracted through natural language processing technology, and the search times of the search terms are recorded; the search terms are matched with the historical browsing text to obtain the historical search terms, and the number of occurrences of the search terms in the historical browsing text is counted, and the number of occurrences and search times of the historical search terms in the historical search text are integrated to obtain the historical search record.

3. The document-assisted reading system based on AI and personalized knowledge base according to claim 2, characterized in that: The step of matching the search terms with the historical browsing text to obtain the historical search terms includes: A1: Extract search terms and historical browsing text; A2: Determine whether the search term appears in the historical browsing text; if yes, mark the corresponding search term as a historical search term; if no, jump to A3; A3: Determine whether the search term has the same semantics as the term in the historical browsing text; if yes, mark the corresponding search term as a historical search term; if no, mark the corresponding search term as an irrelevant term.

4. The document-assisted reading system based on AI and personalized knowledge base according to claim 1, characterized in that: The method of obtaining the user's mastery of the interpreted vocabulary based on the historical search records and the interpreted vocabulary includes: S1: Extract historical search records and interpret vocabulary; S2: Determine whether the historical search words in the historical search records contain the interpretation words; if yes, jump to S3; if no, mark the corresponding interpretation words as new words; S3: Obtaining the mastery status of the interpreted vocabulary according to the number of appearances and the number of searches of the historical search vocabulary corresponding to the interpreted vocabulary.

5. The document-assisted reading system based on AI and personalized knowledge base according to claim 4, characterized in that: The method of obtaining the mastery status of the interpreted vocabulary according to the number of occurrences and the number of searches of the historical search vocabulary corresponding to the interpreted vocabulary includes: B1: Extract the occurrence and search times of historical search terms; B2: Determine whether the number of searches is greater than 1; if yes, jump to B3; if no, jump to B4; B3: Determine whether the number of occurrences is greater than the number of searches; if yes, mark the mastery status of the interpretation vocabulary as mastered; if no, mark the mastery status of the corresponding interpretation vocabulary as difficult to master; B4: Determine whether the number of occurrences is greater than the number of searches; if yes, mark the mastery status of the corresponding interpretation vocabulary as fully mastered; if no, mark the mastery status of the corresponding interpretation vocabulary as mastered; B5: For the interpretation vocabulary whose mastery status is mastered, set the mastery coefficient, and multiply the ratio of the number of appearances to the number of searches by the mastery coefficient to obtain the mastery value of the interpretation vocabulary; wherein the mastery coefficient is set according to the user's ability to understand the vocabulary; B6: Set a mastery threshold; determine whether the mastery value of the interpretation vocabulary is greater than or equal to the mastery threshold; if yes, mark the mastery status of the corresponding interpretation vocabulary as specific mastery; if no, mark the mastery status of the corresponding interpretation vocabulary as rough mastery.

6. The document-assisted reading system based on AI and personalized knowledge base according to claim 5, characterized in that: The mastery coefficient is set according to the user's ability to understand vocabulary, including: Count the historical search words whose appearance times are greater than the search times and mark them as historical mastered words, extract the appearance time and search time of the historical mastered words, and mark the appearance time closest to the current time as the end appearance time, and mark the search time closest to the current time as the end search time; Determine whether the terminal appearance time is greater than the terminal search time; if yes, mark the corresponding historically mastered vocabulary as mastered vocabulary; if no, mark the corresponding historically mastered vocabulary as unmastered vocabulary; The ratio of the mastered vocabulary to the historically mastered vocabulary is marked as the mastery coefficient.

7. The document-assisted reading system based on AI and personalized knowledge base according to claim 1, characterized in that: The auxiliary tags corresponding to the interpretation words generated based on the mastery status are added to the key words, including: Extract mastery status; The interpretation vocabulary that is difficult to master or roughly mastered is marked as auxiliary reading vocabulary, the text type of the reading text is obtained through natural language processing technology, the professional database corresponding to the text type is matched in the database, the auxiliary reading vocabulary is input into the professional database to obtain relevant information corresponding to the auxiliary reading vocabulary, the relevant information is input into the label file to obtain auxiliary tags, and the auxiliary tags are added to the key vocabulary.

Images