A task processing method, device, equipment and medium for smart office

Abstract

The present invention is applicable to the technical field of data identification, and provides a task processing method, device, equipment and medium for smart office, wherein the task processing method for smart office includes: obtaining text data to be identified and primary classification information of the text data to be identified; extracting the current title part and / or the current summary part in the text data to be identified; determining the target secondary classification information of the text data to be identified based on the current title part and / or the current summary part; extracting the text area corresponding to the target key content category in the text data to be identified based on the keywords corresponding to the target key content category; marking the target key content categories corresponding to multiple text areas in the text data to be identified. This technical solution can greatly improve the automation and accuracy of document processing through intelligent document classification, key content extraction and area labeling.

Description

Technical Field

[0001] The present invention belongs to the technical field of data identification, and in particular, relates to a task processing method, device, equipment and medium for smart office. Background Art

[0002] With the advent of the information age, the demand for document processing in modern offices is growing, especially in the processing and information extraction of massive documents. Traditional manual processing methods can no longer meet the requirements of efficiency and accuracy. Many companies and institutions are faced with the problem of how to quickly identify and extract valuable information from complex documents. Especially in document types such as scientific research reports, paper journals, contracts, etc., identifying and extracting key content can not only improve work efficiency, but also reduce human errors and improve the quality of document processing.

[0003] At present, there are some intelligent document analysis and task processing methods based on natural language processing (NLP) technology, which usually assist users to quickly obtain the required information through steps such as text classification, summary extraction, and information retrieval. However, most existing technologies are limited to simple keyword matching or text classification, and lack the ability to accurately locate and annotate the real key content in the document. For documents with different fields, formats, and structures, existing technologies often cannot intelligently identify different categories of key content based on document type, which is a technical problem that needs to be solved urgently. Summary of the invention

[0004] In view of this, an embodiment of the present invention provides a task processing method, device, equipment and medium for smart office to solve the technical problem that the prior art often cannot intelligently identify key content of different categories according to document type.

[0005] A first aspect of an embodiment of the present invention provides a task processing method for a smart office, and the task processing method for a smart office includes:

[0006] Acquire text data to be identified and primary classification information of the text data to be identified; the primary classification information includes research reports, journals or contracts;

[0007] Extracting the current title part and / or the current abstract part in the text data to be recognized;

[0008] Determine target secondary classification information of the text data to be identified according to the current title part and / or the current abstract part; the target secondary classification information is the field category corresponding to each of the primary classification information;

[0009] Acquire a target key content category corresponding to the target secondary classification information and a keyword corresponding to the target key content category; the target key content category is a key content selected by a user from a plurality of preset key content categories; the keyword includes a core word and an associated word;

[0010] Extracting a text area corresponding to the target key content category in the text data to be identified according to the keywords corresponding to the target key content category;

[0011] In the text data to be recognized, target key content categories corresponding to a plurality of the text regions are marked.

[0012] Furthermore, the step of determining target secondary classification information of the text data to be identified according to the current title part and / or the current summary part includes:

[0013] Performing word segmentation processing on the current title part and / or the current abstract part to obtain a plurality of first words;

[0014] Obtain all secondary classification information corresponding to the primary classification information;

[0015] Acquire multiple word sets corresponding to all secondary classification information; wherein each of the word sets includes preset sample words obtained after word segmentation processing of a single sample title part or a single sample summary part;

[0016] For each first word, matching a target word set having the first word;

[0017] The target secondary classification information of the text data to be recognized is determined according to the multiple initial secondary classification information corresponding to each of the multiple target word sets.

[0018] Furthermore, the step of determining target secondary classification information of the text data to be recognized according to the multiple initial secondary classification information corresponding to each of the multiple target word sets includes:

[0019] According to the mapping relationship between the first word, the target word set and the initial secondary classification information, the number of initial secondary classification information mapped to each first word is counted;

[0020] The first word whose amount of initial secondary classification information is less than the first threshold is used as the target first word;

[0021] Counting multiple initial secondary classification information corresponding to all target first words to form a secondary classification information set;

[0022] Counting second quantities corresponding to different types of secondary classification information in the secondary classification information set;

[0023] The secondary classification information corresponding to the maximum second quantity is used as the target secondary classification information.

[0024] Furthermore, the step of extracting the text area corresponding to the target key content category in the text data to be identified according to the keywords corresponding to the target key content category includes:

[0025] Obtaining a preset weight corresponding to the associated word;

[0026] Segment the text data to be recognized to obtain a plurality of sentences to be recognized;

[0027] Extracting a current sentence to be recognized that has the core word from among the multiple sentences to be recognized;

[0028] Extracting the current associated words and the number of the current associated words in each of the current sentences to be recognized;

[0029] Calculating the relevance of the current sentence to be recognized according to the preset weight of the current associated word and the current number of associated words in the current sentence to be recognized;

[0030] The current sentence to be recognized, whose relevance is greater than the second threshold, is used as the text region corresponding to the target key content category.

[0031] Furthermore, the step of calculating the relevance of the current sentence to be recognized according to the preset weight of the current associated words and the current number of associated words in the current sentence to be recognized includes:

[0032] Adding the preset weights of all current associated words in the current sentence to be recognized to obtain a total weight;

[0033] Multiplying the total weight by a first adjustment coefficient to obtain a first sub-score;

[0034] Multiplying the current number of associated words by a second adjustment coefficient to obtain a second sub-score;

[0035] The first sub-score and the second sub-score are added to obtain the association degree.

[0036] Furthermore, before the step of obtaining the preset weight corresponding to the associated word, the method further includes:

[0037] Obtain multiple sample sentences with core words;

[0038] Performing word segmentation processing on the plurality of sample sentences respectively to obtain a plurality of sample words;

[0039] Among all the sample words, obtain the number of occurrences of different sample words to be counted;

[0040] Divide the number of times by the total number of sample words to obtain the word frequency;

[0041] Counting target sample sentences having the sample words to be counted;

[0042] Dividing the total number of sample sentences by the target number of sample sentences to obtain a first value;

[0043] Calculating the logarithm of the first value to obtain a rarity index of the sample word to be counted;

[0044] The word frequency corresponding to each sample word to be counted is multiplied by the rarity index to obtain the preset weights corresponding to different sample words to be counted.

[0045] Furthermore, the step of multiplying the word frequency corresponding to each sample word to be counted by the rarity index to obtain the preset weights corresponding to different sample words to be counted includes:

[0046] Multiply the word frequency and rarity index of each sample word to be counted to obtain the initial weight;

[0047] When the initial weight of the sample word pair to be counted is greater than the third threshold, the initial weight is used as the preset weight corresponding to the sample word to be counted;

[0048] When the initial weight of the sample word pair to be counted is not greater than the third threshold, the sample word to be counted is eliminated.

[0049] A second aspect of an embodiment of the present invention provides a task processing device for smart office, including:

[0050] A first acquisition unit is used to acquire text data to be identified and primary classification information of the text data to be identified; the primary classification information includes research reports, paper journals or contracts;

[0051] A first extraction unit, used for extracting a current title part and / or a current abstract part in the text data to be recognized;

[0052] A determination unit, configured to determine target secondary classification information of the text data to be identified according to the current title part and / or the current abstract part; the target secondary classification information is the field category corresponding to each of the primary classification information;

[0053] The second acquisition unit is used to acquire a target key content category corresponding to the target secondary classification information and a keyword corresponding to the target key content category; the target key content category is a key content selected by a user from a plurality of preset key content categories; the keyword includes a core word and an associated word;

[0054] A second extraction unit, configured to extract a text region corresponding to the target key content category in the text data to be identified according to the keywords corresponding to the target key content category;

[0055] The marking unit is used to mark the target key content categories corresponding to the multiple text areas in the text data to be recognized.

[0056] A third aspect of an embodiment of the present invention provides a terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of the task processing method for smart office described in the first aspect are implemented.

[0057] A fourth aspect of an embodiment of the present invention provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps of the task processing method of smart office described in the first aspect are implemented.

[0058] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: by obtaining the primary classification information of the text data to be identified (such as research reports, journals or contracts, etc.), and further analyzing the secondary classification information based on the information, the specific category of the document can be accurately identified according to the characteristics of different document types, avoiding the fuzzy or wrong classification problems in traditional methods. By combining the title and abstract parts of the document for analysis, the accuracy of classification is effectively improved, especially when processing complex or multi-category documents, the subdivision of the document can be automatically identified.

[0059] After determining the target secondary classification of the document, this solution further extracts the corresponding key content categories based on the target secondary classification information. By matching with multiple preset key content categories and combining with the keywords selected by the user (including core words and associated words), the key content areas in the document can be automatically located and accurately extracted. This keyword-based intelligent extraction method not only improves the extraction efficiency, but also ensures the relevance and accuracy of the extraction results, avoiding the inefficiency and errors of manual extraction.

[0060] This solution provides highly automated support for document analysis, processing, and information retrieval by marking text areas corresponding to multiple target key content categories. Users do not need to manually mark key content, and the system automatically identifies and marks key content areas in the document, greatly improving work efficiency. It is especially suitable for scenarios that require processing a large number of documents, such as legal documents, scientific research papers, and other professional documents.

[0061] This solution is applicable to many different types of documents, especially research reports, journals, contracts and other document categories, and is highly universal. Users can flexibly adjust key content categories and keywords according to different needs and fields to meet the needs of intelligent recognition in different scenarios, and it has wide application potential.

[0062] By classifying and extracting target key content based on document structure (such as title and abstract), this technology can effectively improve the efficiency and accuracy of subsequent information retrieval. Through automated annotation and extraction, users can quickly locate relevant key information and avoid interference from redundant and irrelevant content.

[0063] In general, this technical solution can greatly improve the automation and accuracy of document processing through intelligent document classification, key content extraction and area labeling, save labor costs and improve work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments or related technical descriptions. 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 creative work.

[0065] Figure 1 A schematic flow chart of a task processing method for smart office provided by the present invention is shown;

[0066] Figure 2 A schematic diagram of a task processing device for smart office provided by an embodiment of the present invention is shown;

[0067] Figure 3 A schematic diagram of a terminal device provided by an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0068] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present invention. However, it should be clear to those skilled in the art that the present invention may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present invention.

[0069] The embodiments of the present invention provide a task processing method, device, equipment and medium for smart office to solve the technical problem that the prior art often cannot intelligently identify key contents of different categories according to document types.

[0070] First, the present invention provides a method, device, equipment and medium for processing tasks in a smart office. Figure 1 , Figure 1 FIG. 1 is a schematic flow chart of a task processing method for a smart office provided by the present invention. Figure 1 As shown, the task processing method of the smart office may include the following steps:

[0071] Step 101: obtaining text data to be identified and primary classification information of the text data to be identified; the primary classification information includes research reports, paper journals or contracts;

[0072] First, obtain the text data to be recognized, usually electronic documents or text files. These texts can be any type of documents. Before entering the recognition process, each document will be classified and stored (such as research reports, papers, journals, or contracts), that is, the first classification information is inherent known information. Among them, the first-level classification information includes but is not limited to research reports, papers, journals, or contracts.

[0073] Step 102: extracting the current title part and / or the current abstract part in the text data to be recognized;

[0074] Extract the title and abstract from the text data to be identified. The title and abstract can usually express the core theme and purpose of the document concisely, so they are important bases for classification and content identification.

[0075] Title section: usually contains the core topic of the document, helping the system quickly understand the overall content of the document.

[0076] Summary: Provides a brief overview or key points of the document to help understand the specific content and background of the document.

[0077] Step 103: determining target secondary classification information of the text data to be identified according to the current title part and / or the current abstract part; the target secondary classification information is the field category corresponding to each of the primary classification information;

[0078] Based on the information in the title and abstract, the system will further subdivide the category of the document. By analyzing the title and abstract, the system can assign a more accurate secondary classification information to the document, thus laying the foundation for subsequent key content identification. Among them, the target secondary classification information of the research report includes but is not limited to market research reports, industry research reports, product research reports, policy research reports or feasibility research reports, the target secondary classification information of paper journals includes but is not limited to natural science paper journals, engineering technology paper journals, medical paper journals and social science paper journals, and the target secondary classification information of the contract includes but is not limited to sales contracts, lease contracts, commission contracts, loan contracts or labor contracts. Among them, the specific classification logic of the target secondary classification information is as follows:

[0079] Specifically, step 103 specifically includes steps 1031 to 1035:

[0080] Step 1031: performing word segmentation processing on the current title part and / or the current abstract part to obtain a plurality of first words;

[0081] Word segmentation is a basic step in natural language processing, which aims to split the text into words for subsequent processing and analysis. In this step, the system first performs word segmentation on the title part and / or abstract part of the document to be recognized.

[0082] Title part: For example, if the title of the document is "2024 Market Trend Analysis Report", after word segmentation, the word set may be obtained: [2024, market, trend, analysis, report].

[0083] Abstract: The abstract of a document may include "This article mainly analyzes the trend changes in the global market in 2024, and makes a detailed market forecast based on the data." After word segmentation, the resulting word set is: [this article, mainly, analysis, global, market, trend, change, data, market, forecast].

[0084] After word segmentation, the result is a set of multiple "first words" (for example, "market", "trend", etc.), which are the basis for subsequent classification analysis.

[0085] Step 1032: Acquire all secondary classification information corresponding to the primary classification information;

[0086] The first-level classification is a large category of documents, such as "research report", "paper journal", "contract", etc. Based on different first-level classifications, the system will define a set of second-level classifications. For example, "research report" may have second-level classifications such as "market research", "industry analysis", etc., while "paper journal" may include "medical paper", "computer science paper", etc. In this step, the system will obtain all the second-level classification information corresponding to the current first-level classification information to provide a reference for subsequent classification matching.

[0087] Step 1033: Acquire multiple word sets corresponding to all secondary classification information; wherein each of the word sets includes preset sample words obtained after word segmentation processing of a single sample title part or a single sample summary part;

[0088] Each secondary classification (such as "market research", "medical papers", etc.) corresponds to a set of keywords or common words. These word sets contain common words or typical terms in the classification field, which help determine the accurate classification of documents. For example, the word set corresponding to the secondary classification "market research" may include: [market, trend, competition, analysis, data], while the word set corresponding to "medical papers" may include: [clinical, experimental, pathology, research, medicine]. These word sets are automatically extracted from a large number of document samples as representative words for each secondary classification.

[0089] Step 1034: For each first word, match a target word set having the first word;

[0090] The system will check the "first words" obtained in step 1031 (for example, the words obtained after the title or abstract is segmented) one by one, and match them with the word sets of all secondary classification information. That is, the system will check which word sets of the secondary classification contain the current "first word".

[0091] Step 1035: Determine the target secondary classification information of the text data to be recognized according to the multiple initial secondary classification information corresponding to each of the multiple target word sets.

[0092] Once the matching is completed, the system will determine the target secondary classification of the document to be identified based on the secondary classification information in the multiple target word sets matched by each first word. The specific logic is as follows:

[0093] Specifically, step 1035 includes steps A1 to A5:

[0094] Step A1: according to the mapping relationship between the first word, the target word set and the initial secondary classification information, counting the number of initial secondary classification information mapped to each first word;

[0095] For each "first term" in the text to be recognized (usually words in the title or abstract), the system looks for its relationships with different "sets of target terms" and corresponding "initial secondary classification information". The system counts the number of "initial secondary classification information" mapped to by each first term. This number indicates the degree of relevance of the term to how many secondary classifications. For example, if the word "market" appears in two categories, "market research" and "market trend", then the number of initial secondary classification information corresponding to "market" is 2.

[0096] Step A2: Use the first terms with the number of initial secondary classification information less than the first threshold as the target first terms;

[0097] Since the first terms with high frequencies may be stop words (such as "of", "is", etc.), although these words have a high word frequency in each document, they have low discriminability for text content. Therefore, these ultra-high-frequency words need to be filtered out through the first threshold.

[0098] By using logarithms, the weights of common words can be effectively reduced, avoiding their excessive influence on the calculation results, so that when distinguishing documents, more attention is paid to rare words that can provide more information.

[0099] This step ensures that the system can focus on those terms with strong classification indicators and avoids the interference of ultra-high-frequency words such as stop words or general words on the classification process. This screening mechanism effectively reduces unnecessary information noise and improves the classification accuracy of the system.

[0100] Step A3: Count the set of secondary classification information formed by multiple initial secondary classification information corresponding to all target first terms;

[0101] Step A4: Count the second quantity corresponding to each different type of secondary classification information in the set of secondary classification information;

[0102] In this step, the system counts the number of times each different type of secondary classification information appears (i.e., the second quantity) in the set of secondary classification information corresponding to the target first terms. This count can reflect the frequency of certain secondary classifications in the set. For example, if "market research" appears frequently among multiple target first terms, while "market trend" appears less frequently, then "market research" will have a higher "second quantity". The second quantity refers to the frequency of each secondary classification information.

[0103] Step A5: Use the secondary classification information corresponding to the largest second quantity as the target secondary classification information.

[0104] According to the second number obtained in step A4, the system selects the second-level classification information with the largest second number as the final target second-level classification information. In other words, the system selects the second-level classification with the highest frequency as the classification result of the document.

[0105] For example, if in a set of classification information of multiple target first words, "market research" appears the most times (for example, 4 times, while "market trend" only appears 3 times), then "market research" will be selected as the final target secondary classification information.

[0106] Through refined word mapping statistics and automatic screening mechanisms, this technical solution can intelligently and accurately determine the target secondary classification information of the text data to be identified. It not only improves classification accuracy and reduces manual intervention, but also has significant efficiency advantages when processing large-scale documents.

[0107] In this embodiment, by performing word segmentation processing on the title part and / or abstract part of the text to be recognized, the text is first divided into multiple words (first words). This word segmentation processing can effectively extract the basic semantic units of the text and help the subsequent classification and matching work. This processing step avoids the complexity of directly processing the original text, improves the operability of the text data, and provides an accurate vocabulary basis for the identification of the target secondary classification information in the subsequent steps. In order to better identify the target secondary classification information, this scheme relies on multiple word sets corresponding to each secondary classification information, wherein these word sets are obtained by performing word segmentation processing on the title or abstract part of the historical sample. Each word set contains preset sample words related to a specific secondary classification and has a high classification representativeness. By matching the words in the current text to be recognized with these word sets, this technology can accurately infer the secondary classification information of the text to be recognized based on the accumulation of historical data and samples. In the matching process, for each first word obtained after the word segmentation processing, the system automatically matches the target word set with the corresponding first word. The matching process utilizes the correlation between words, ensures the accurate positioning of the target word set, and enhances the accuracy of classification by matching multiple target word sets. This process can help the system quickly identify and locate the domain category to which the document belongs, avoiding the subjectivity and errors of manual classification. After completing the matching of multiple target word sets, this solution will comprehensively analyze and determine the final target secondary classification information of the text data to be identified based on the multiple initial secondary classification information corresponding to each word set. This comprehensive analysis step not only considers the matching results of a single word, but also improves the reliability and accuracy of the classification through the verification of multiple word sets, ensuring that the system can make scientific and reasonable judgments on the secondary classification of different types of documents. Through the implementation of this technical solution, the accuracy and automation level of document classification can be effectively improved. The system can automatically determine the domain category of the document based on the word segmentation results of the current title part and the abstract part, avoiding the human factors and cumbersome processes in traditional manual or semi-automatic classification methods. This method is particularly suitable for large-scale document processing scenarios, such as automatic classification of scientific research documents, intelligent archiving of legal documents, etc., significantly reducing manual intervention and improving classification efficiency. This technical solution can not only meet the classification needs of various document types (such as scientific research papers, research reports, legal contracts, etc.), but also has strong scalability. Users can flexibly adjust sample word sets or add new domain categories according to actual application needs, thereby ensuring that the system can cope with ever-changing document classification tasks. Overall, this technical solution can achieve fast and efficient secondary classification information recognition through innovative text segmentation processing and target word set matching mechanism, greatly improving the accuracy and automation level of text classification and adapting to large-scale document processing needs.

[0108] Step 104: obtaining a target key content category corresponding to the target secondary classification information and a keyword corresponding to the target key content category; the target key content category is a key content selected by the user from a plurality of preset key content categories; the keyword includes a core word and an associated word;

[0109] The system displays optional preset key content categories to the user through the display module of the terminal, and the user selects a target key content category from the optional preset key content categories, and then the system identifies and marks the corresponding text content according to the target key content category selected by the user.

[0110] Each secondary classification information (such as market research, medical papers, sales contracts, etc.) corresponds to a certain key content category. The key content category is pre-set and usually covers the most representative content in the document. For example, a market research report may include key content categories such as "market trends", "competitive analysis", and "customer needs". Medical papers may include "experimental methods", "results analysis", "clinical trials", etc.

[0111] Each target key content category will have a set of keywords, which include core words and associated words: Core words: point to the most important words in the category, which can directly express the theme or core idea of ​​the category. For example, in the "Market Trends" category, core words may be "growth", "trends", "market demand", etc. Associated words: secondary words related to core words, which help improve the accuracy of text understanding. For example, the "Competitive Analysis" category may contain words such as "market share" or "SWOT analysis".

[0112] Step 105: extracting a text region corresponding to the target key content category in the text data to be identified according to the keywords corresponding to the target key content category;

[0113] According to the keywords determined in step 104, the system will scan the text data to be recognized, locate and extract text areas related to these keywords. These text areas contain key information of the category.

[0114] Specifically, step 105 specifically includes steps 1051 to 1056:

[0115] Step 1051: Obtaining the preset weight corresponding to the associated word;

[0116] Related words are words related to a specific "target key content category" in the text. These related words are key clues used to help determine which parts of the text are related to the target category. For example, assuming the target key content category is "market trends", related words may include "growth" or "forecast".

[0117] Each associated word will have a preset weight, and the size of the weight indicates the importance of the associated word in representing the target content category. For example, "growth" may have a high weight, while "forecast" may have a medium weight. The preset weight is set in advance and is used to calculate the relevance of the statement later.

[0118] Get the preset weights of the associated words. By assigning a weight value to each associated word, the system can distinguish different words according to their importance in the text. The preset weight of the associated word represents the closeness of the word to the target key content category. The higher the weight value, the stronger the relevance of the associated word to the target content. In this way, the system can finely adjust the influence of different associated words on the target content extraction process to ensure the accuracy and representativeness of the extraction results. The calculation logic of the preset weights is as follows:

[0119] Optionally, before step 1051, steps C1 to C8 are also included:

[0120] Step C1: obtaining a plurality of sample sentences having core words;

[0121] Step C2: performing word segmentation processing on the plurality of sample sentences respectively to obtain a plurality of sample words;

[0122] Step C3: among all sample words, obtain the number of occurrences of different sample words to be counted;

[0123] Step C4: Divide the number of times by the total number of sample words to obtain the word frequency;

[0124] The frequency of a word refers to the number of times a particular word appears in all sample sentences, divided by the total number of words (i.e., the total number of words in all sample sentences).

[0125] Step C5: Counting target sample sentences having the sample words to be counted;

[0126] Step C6: Divide the total number of sample sentences by the target number of sample sentences to obtain a first value;

[0127] The first value is calculated by dividing the total number of sample sentences by the number of target sample sentences containing the specific word to be counted. This value measures the extent to which the word appears in the sample sentences.

[0128] Step C7: Calculate the logarithm of the first value to obtain the rarity index of the sample word to be counted;

[0129] Rarity reflects the distribution of a word in the corpus. A lower first value indicates that the word appears in relatively few sentences, so it may be a rarer word. By taking the logarithm of the first value, the rarity index of the word can be obtained. The calculation of the logarithm helps to reduce the impact of frequently occurring words on the rarity score (because frequently occurring words are usually stop words or common words, they have a certain impact on recognition and classification, so the impact of these words needs to be reduced), and the logarithm calculation can suppress the impact of ultra-high frequency words, so that those rarer words can be assigned higher rarity values.

[0130] Step C8: multiply the word frequency corresponding to each sample word to be counted by the rarity index to obtain the preset weight corresponding to each sample word to be counted.

[0131] Specifically, step C8 specifically includes steps C81 to C83:

[0132] Step C81: multiply the word frequency corresponding to each sample word to be counted by the rarity index to obtain an initial weight;

[0133] Step C82: when the initial weight of the sample word pair to be counted is greater than the third threshold, the initial weight is used as the preset weight corresponding to the sample word to be counted;

[0134] The third threshold is a preset value used to distinguish which words have sufficient weight for subsequent processing. This threshold is set according to business needs, data characteristics or application scenarios. It is usually used to filter out some unimportant or uncommon words.

[0135] When the initial weight of a word is greater than this threshold, it means that the word has a high frequency and moderate rarity in the corpus, which means that the word is relatively important and can be used as an effective associated word. Therefore, the system will retain this initial weight as the preset weight of the word.

[0136] If the initial weight of a word is greater than the third threshold, the preset weight of the word is equal to its initial weight. For example, if the initial weight of a word is 0.014, and the third threshold is set to 0.01, then since 0.014>0.01, the word will retain this initial weight as the preset weight.

[0137] Step C83: When the initial weight of the sample word pair to be counted is not greater than the third threshold, the sample word to be counted is eliminated.

[0138] When the initial weight of a word is less than or equal to the third threshold, it means that the word has a low relevance, which may be due to the low frequency of occurrence or low rarity of the word, and therefore has a small contribution to the target task. In order to avoid introducing noise, the system removes the word from the sample to be counted.

[0139] The purpose of this step is to remove some words with low initial weights to reduce the impact of irrelevant or insignificant words on the analysis results. After removal, these words will no longer participate in the subsequent preset weight calculation or relevance evaluation. For example, if the initial weight of a word is 0.005, and the third threshold is set to 0.01, then since 0.005<0.01, the word will be removed and will no longer be counted.

[0140] In this way, the system can effectively filter out words that are of low frequency, too rare, or have little contribution to the task, retaining only important words and assigning them appropriate weights.

[0141] In this embodiment, the preset weight of each sample word to be counted is calculated by combining the word frequency and rarity index, ensuring that the frequency of word occurrence and its rarity can be taken into account in text analysis. Through this comprehensive approach, the system can dynamically adjust the importance of each word in the target text, thereby more accurately reflecting the actual influence of the word. By combining frequency and rarity, the system can more flexibly weigh the importance of different words in the analysis. In traditional text analysis methods, word frequency is often emphasized, while the importance of low-frequency words is ignored. This solution can overcome this shortcoming, so that low-frequency but important words can get higher weights, thereby improving the accuracy and comprehensiveness of the analysis. In summary, the preset weights calculated by the combination of word frequency and rarity in this technical solution can effectively improve the accuracy and intelligence level of text analysis.

[0142] As an optional embodiment of the present invention, the preset weight can also be calculated by the following function:

[0143] A plurality of sample sentences having core words are obtained, and the plurality of sample sentences are respectively segmented to obtain a plurality of sample words.

[0144]

[0145] in, represents the preset weight of the jth associated word, represents the jth associated word, Indicates the jth associated word The frequency of occurrence of all sample words, represents the number of all sample words, represents the number of sample sentences, Represents an indicator function, which takes the value 1 when the jth associated word appears in the current sample sentence, otherwise it takes the value 0. Represents the i-th sample sentence. , and Indicates the conversion factor.

[0146] in, By calculating the frequency of each word in a single sentence, the relative importance of the word in the sentence is measured. Words with high frequency contribute more to the content of the sentence and are therefore often considered more important.

[0147] It measures the distribution (breadth of distribution) of the jth associated word in the entire sample sentence. A word that appears in multiple sentences may be more universal in semantic analysis because it can span multiple contexts. Therefore, the higher the coverage, the wider the distribution of the word and the greater its overall influence on the entire sample sentence.

[0148] The rarity of the jth associated word in the corpus is calculated. Specifically, rarity is defined by the inverse of the total number of occurrences of the word, and logarithmic smoothing is used to avoid extreme values. Frequently occurring words (such as common function words and stop words) are not very recognizable in the corpus, so their weights should be relatively low. Conversely, rare words have a greater unique contribution to the corpus, so words with higher rarity are usually given higher weights. The purpose of logarithmic smoothing is to prevent frequently occurring words from having too high rarity values ​​and ensure the stability of the calculation.

[0149] Through these comprehensive calculations, we finally get the preset weight of each target word. This weight takes into account the frequency of the word, as well as its scarcity, coverage, and relationship with the target word. The core principle of the above function is to evaluate the "importance" of each word from multiple angles, including its commonness (frequency), distribution breadth (coverage), and scarcity (rarity). , and The three result values ​​are mapped and transformed. This design can more accurately evaluate the relative weight of words.

[0150] Step 1052: Segment the text data to be recognized to obtain a plurality of sentences to be recognized;

[0151] Text data (such as articles, reports, news, etc.) are first processed into multiple independent sentences. Each sentence is usually a relatively complete unit of thought, which may be a sentence, phrase or paragraph.

[0152] In order to better analyze the text data to be recognized, the solution first divides the text into sentences to obtain multiple sentences to be recognized. This step divides large sections of text into structured small blocks (i.e. sentences), which facilitates the subsequent analysis and processing of the content of each sentence. Sentence processing not only improves the efficiency of text analysis, but also ensures that the extraction process can be carried out within a smaller context, thereby improving the extraction accuracy.

[0153] Step 1053: extracting the current sentence to be recognized that has the core word from the multiple sentences to be recognized;

[0154] The system looks for core words in the text to be identified (for example, words related to the target key content category). These core words may be "market", "growth", "trend", etc.

[0155] After obtaining multiple sentences to be identified, the system will automatically extract the current sentence to be identified that contains the target core word. Core words are usually the core identifiers of the target key content category, and sentences containing core words can often more accurately reflect the key content of the category. By identifying core words, the system can focus on the text area that is most likely to contain the target key information, avoiding interference from irrelevant content and improving the relevance of extraction.

[0156] Step 1054: extracting the current associated words and the number of current associated words in each of the current sentences to be recognized;

[0157] For each sentence to be identified, the system checks whether it contains the defined associated words. For example, if the target key content category is "market trends" and the associated words include "growth", "forecast", "market", the system checks whether the current sentence contains these words.

[0158] Step 1055: Calculate the relevance of the current sentence to be recognized according to the preset weight of the current associated word and the current number of associated words in the current sentence to be recognized;

[0159] Based on this information, combined with the preset weights of the associated words, the system calculates the relevance of each current sentence to be identified. The calculation of the relevance takes into account the number of associated words and their weight values, ensuring that more frequently occurring and highly weighted associated words have a greater impact on the final result. In this way, the system can more intelligently evaluate the relevance of each sentence to the target key content category.

[0160] Specifically, step 1055 specifically includes steps B1 to B4:

[0161] Step B1: adding the preset weights of all current associated words in the current sentence to be recognized to obtain a total weight;

[0162] For the current sentence to be recognized, the system will list all the associated words that appear in the sentence and add up the preset weights of these associated words.

[0163] Step B2: multiplying the total weight by the first adjustment coefficient to obtain a first sub-score;

[0164] The first adjustment coefficient is a preset factor used to adjust the impact of the total weight on the relevance score. For example, this coefficient can be a constant, or dynamically adjusted according to different application scenarios (such as document type, corpus characteristics).

[0165] Step B3: multiplying the current number of associated words by a second adjustment coefficient to obtain a second sub-score;

[0166] The second adjustment factor is another adjustment factor used to control the effect of the number of associated words on the score.

[0167] Step B4: Add the first sub-score and the second sub-score to obtain the degree of association.

[0168] The final relevance is the sum of the first sub-score and the second sub-score. This relevance value represents the relevance score of the current sentence to be identified with the target content category. A higher relevance means that the sentence is more relevant to the target content category, while a lower relevance means that the relevance is weaker.

[0169] In this embodiment, the preset weight of the current associated words and the current number of associated words are processed separately, and combined with the adjustment coefficient, the relevance of the sentence and the target content is comprehensively evaluated. In this way, the system not only considers the importance of the associated words (reflected by the weight), but also comprehensively considers the frequency of occurrence of the associated words in the sentence (reflected by the number), so that the relevance of the sentence can be comprehensively and accurately evaluated, avoiding the deviation that may be caused by a single dimensional indicator.

[0170] Step 1056: The current sentence to be recognized, whose relevance is greater than the second threshold, is used as the text region corresponding to the target key content category.

[0171] In this step, the system sets a second threshold to filter out sentences with low relevance to the target key content category. This threshold is usually a preset value, which indicates the minimum relevance a sentence needs to reach in order to be considered as a valid text area containing the target key content category.

[0172] By calculating the relevance of each sentence to be identified, the system will automatically select sentences with a relevance greater than the second threshold as the text area of ​​the target key content category. The second threshold can be adjusted according to actual application needs to ensure that only those sentences with high relevance are selected into the final key content area. This screening process effectively reduces the interference of low-relevance sentences on the classification results, ensuring that the final extracted text area is highly relevant and meets the extraction requirements of the target key content category.

[0173] In this embodiment, through the above steps, the system can efficiently and accurately extract the text area corresponding to the target key content category in the text data to be identified, which greatly improves the accuracy and automation level of content extraction. Compared with traditional manual screening or methods based on simple keyword matching, this solution can better capture the potential important information area in the text, avoid the interference of irrelevant information, and reduce the cost of manual intervention. This technical solution can be widely used in various fields that require intelligent extraction of key content, such as legal document analysis, scientific research document processing, technical report review, etc. Whether it is processing legal clauses in contracts, experimental results in scientific research papers, or conclusions in reports, the system can automatically extract relevant text areas according to the target key content category to meet the content extraction needs in different scenarios. In summary, this technical solution can efficiently and accurately extract text areas related to the target key content category in the text data to be identified through intelligent analysis, weight calculation and correlation screening of core words and associated words, greatly improving the accuracy and automation of content extraction.

[0174] Step 106: In the text data to be recognized, mark the target key content categories corresponding to the multiple text regions.

[0175] The system will mark the text areas related to the target key content category in the text data to be identified. These annotations can be done in different ways, such as highlighting, adding labels, etc., to facilitate subsequent manual or automatic processing.

[0176] In this embodiment, by obtaining the primary classification information of the text data to be identified (such as research reports, journals or contracts, etc.), and further analyzing the secondary classification information based on this information, the specific category of the document can be accurately identified according to the characteristics of different document types, avoiding the fuzzy or wrong classification problems in traditional methods. By combining the title and abstract parts of the document for analysis, the accuracy of classification is effectively improved, especially when dealing with complex or multi-category documents, the subdivision of the document can be automatically identified.

[0177] After determining the target secondary classification of the document, this solution further extracts the corresponding key content categories based on the target secondary classification information. By matching with multiple preset key content categories and combining with the keywords selected by the user (including core words and associated words), the key content areas in the document can be automatically located and accurately extracted. This keyword-based intelligent extraction method not only improves the extraction efficiency, but also ensures the relevance and accuracy of the extraction results, avoiding the inefficiency and errors of manual extraction.

[0178] This solution provides highly automated support for document analysis, processing, and information retrieval by marking text areas corresponding to multiple target key content categories. Users do not need to manually mark key content, and the system automatically identifies and marks key content areas in the document, greatly improving work efficiency. It is especially suitable for scenarios that require processing a large number of documents, such as legal documents, scientific research papers, and other professional documents.

[0179] This solution is applicable to many different types of documents, especially research reports, journals, contracts and other document categories, and is highly universal. Users can flexibly adjust key content categories and keywords according to different needs and fields to meet the needs of intelligent recognition in different scenarios, and it has wide application potential.

[0180] By classifying and extracting target key content based on document structure (such as title and abstract), this technology can effectively improve the efficiency and accuracy of subsequent information retrieval. Through automated annotation and extraction, users can quickly locate relevant key information and avoid interference from redundant and irrelevant content.

[0181] In general, this technical solution can greatly improve the automation and accuracy of document processing through intelligent document classification, key content extraction and area labeling, save labor costs and improve work efficiency.

[0182] like Figure 2 The present invention provides a task processing device for smart office, see Figure 2 , Figure 2 A schematic diagram of a task processing device for smart office provided by the present invention is shown. Figure 2 A task processing device for a smart office includes:

[0183] The first acquisition unit 21 is used to acquire the text data to be identified and the primary classification information of the text data to be identified; the primary classification information includes research reports, paper journals or contracts;

[0184] A first extraction unit 22, used for extracting a current title part and / or a current abstract part in the text data to be recognized;

[0185] A determination unit 23 is used to determine target secondary classification information of the text data to be identified according to the current title part and / or the current abstract part; the target secondary classification information is the field category corresponding to each of the primary classification information;

[0186] The second acquisition unit 24 is used to acquire a target key content category corresponding to the target secondary classification information and a keyword corresponding to the target key content category; the target key content category is a key content selected by a user from a plurality of preset key content categories; the keyword includes a core word and an associated word;

[0187] A second extraction unit 25 is used to extract a text area corresponding to the target key content category in the text data to be identified according to the keywords corresponding to the target key content category;

[0188] The marking unit 26 is used to mark the target key content categories corresponding to the multiple text regions in the text data to be recognized.

[0189] The task processing device for smart office provided by the present invention can accurately identify the specific category of the document according to the characteristics of different document types by obtaining the primary classification information of the text data to be identified (such as research reports, journals or contracts, etc.), and further analyzing the secondary classification information based on the information, thereby avoiding the fuzzy or wrong classification problems in traditional methods. By combining the title and abstract parts of the document for analysis, the accuracy of classification is effectively improved, especially when processing complex or multi-category documents, the subdivision of the document can be automatically identified.

[0190] After determining the target secondary classification of the document, this solution further extracts the corresponding key content categories based on the target secondary classification information. By matching with multiple preset key content categories and combining with the keywords selected by the user (including core words and associated words), the key content areas in the document can be automatically located and accurately extracted. This keyword-based intelligent extraction method not only improves the extraction efficiency, but also ensures the relevance and accuracy of the extraction results, avoiding the inefficiency and errors of manual extraction.

[0191] This solution provides highly automated support for document analysis, processing, and information retrieval by marking text areas corresponding to multiple target key content categories. Users do not need to manually mark key content, and the system automatically identifies and marks key content areas in the document, greatly improving work efficiency. It is especially suitable for scenarios that require processing a large number of documents, such as legal documents, scientific research papers, and other professional documents.

[0192] This solution is applicable to many different types of documents, especially research reports, journals, contracts and other document categories, and is highly universal. Users can flexibly adjust key content categories and keywords according to different needs and fields to meet the needs of intelligent recognition in different scenarios, and it has wide application potential.

[0193] By classifying and extracting target key content based on document structure (such as title and abstract), this technology can effectively improve the efficiency and accuracy of subsequent information retrieval. Through automated annotation and extraction, users can quickly locate relevant key information and avoid interference from redundant and irrelevant content.

[0194] In general, this technical solution can greatly improve the automation and accuracy of document processing through intelligent document classification, key content extraction and area labeling, save labor costs and improve work efficiency.

[0195] Figure 3 Schematic diagram of a terminal device provided by an embodiment of the present invention. Figure 3 As shown, a terminal device 3 of this embodiment includes: a processor 30, a memory 31, and a computer program 32 stored in the memory 31 and executable on the processor 30, such as a task processing program for smart office. When the processor 30 executes the computer program 32, the steps in each of the above-mentioned smart office task processing method embodiments are implemented, such as Figure 1 Alternatively, when the processor 30 executes the computer program 32, the functions of each unit in the above-mentioned device embodiments are realized, for example, Figure 2 Function of the unit shown.

[0196] Exemplarily, the computer program 32 may be divided into one or more units, which are stored in the memory 31 and executed by the processor 30 to complete the present invention. The one or more units may be a series of computer program instruction segments capable of completing specific functions, which are used to describe the execution process of the computer program 32 in the terminal device 3. For example, the computer program 32 may be divided into the following specific functions of each unit:

[0197] A first acquisition unit is used to acquire text data to be identified and primary classification information of the text data to be identified; the primary classification information includes research reports, paper journals or contracts;

[0198] A first extraction unit, used for extracting a current title part and / or a current abstract part in the text data to be recognized;

[0199] A determination unit, configured to determine target secondary classification information of the text data to be identified according to the current title part and / or the current abstract part; the target secondary classification information is the field category corresponding to each of the primary classification information;

[0200] The second acquisition unit is used to acquire a target key content category corresponding to the target secondary classification information and a keyword corresponding to the target key content category; the target key content category is a key content selected by a user from a plurality of preset key content categories; the keyword includes a core word and an associated word;

[0201] A second extraction unit, configured to extract a text region corresponding to the target key content category in the text data to be identified according to the keywords corresponding to the target key content category;

[0202] The marking unit is used to mark the target key content categories corresponding to the multiple text areas in the text data to be recognized.

[0203] The terminal device includes but is not limited to a processor 30 and a memory 31. Those skilled in the art will appreciate that Figure 3 It is only an example of a terminal device 3 and does not constitute a limitation on the terminal device 3. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the terminal device may also include input and output devices, network access devices, buses, etc.

[0204] The processor 30 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0205] The memory 31 may be an internal storage unit of the terminal device 3, such as a hard disk or memory of the terminal device 3. The memory 31 may also be an external storage device of the terminal device 3, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the terminal device 3. Further, the memory 31 may also include both an internal storage unit and an external storage device of the terminal device 3. The memory 31 is used to store the computer program and other programs and data required by the roaming control device. The memory 31 may also be used to temporarily store data that has been output or is to be output.

[0206] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0207] It should be noted that the information interaction, execution process, etc. between the above-mentioned devices / units are based on the same concept as the method embodiment of the present invention. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.

[0208] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In actual applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated into a processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0209] An embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the above-mentioned method embodiments can be implemented.

[0210] An embodiment of the present invention provides a computer program product. When the computer program product runs on a mobile terminal, the mobile terminal can implement the steps in the above-mentioned method embodiments when executing the computer program product.

[0211] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, which can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may at least include: any entity or device that can carry the computer program code to the camera device / terminal device, recording medium, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium. For example, a USB flash drive, a mobile hard disk, a disk or an optical disk.

[0212] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0213] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0214] In the embodiments provided by the present invention, it should be understood that the disclosed devices / network equipment and methods can be implemented in other ways. For example, the device / network equipment embodiments described above are only schematic. For example, the division of the modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0215] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed over multiple network units.

[0216] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0217] It should also be understood that the term "and / or" used in the present description and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0218] As used in the present specification and the appended claims, the term "if" can be interpreted as "when" or "uponce" or "in response to determining" or "in response to monitoring, depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is monitored" can be interpreted as meaning "uponce it is determined" or "in response to determining" or "uponce [described condition or event] is monitored" or "in response to monitoring [described condition or event]", depending on the context.

[0219] In addition, in the description of the present specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0220] References to "one embodiment" or "some embodiments" etc. described in the present specification mean that one or more embodiments of the present invention include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0221] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A task processing method for smart office, characterized in that: The task processing method of the smart office includes: Acquire text data to be identified and primary classification information of the text data to be identified; the primary classification information includes research reports, journals or contracts; Extracting the current title part and / or the current abstract part in the text data to be recognized; Performing word segmentation processing on the current title part and / or the current abstract part to obtain a plurality of first words; Obtain all secondary classification information corresponding to the primary classification information; Acquire multiple word sets corresponding to all secondary classification information; wherein each of the word sets includes preset sample words obtained after word segmentation processing of a single sample title part or a single sample summary part; For each first word, matching a target word set having the first word; Determine the target secondary classification information of the text data to be recognized according to the multiple initial secondary classification information corresponding to each of the multiple target word sets; the target secondary classification information is the field category corresponding to each of the primary classification information; Acquire a target key content category corresponding to the target secondary classification information and a keyword corresponding to the target key content category; the target key content category is a key content selected by a user from a plurality of preset key content categories; the keyword includes a core word and an associated word; Obtaining a preset weight corresponding to the associated word; Segment the text data to be recognized to obtain a plurality of sentences to be recognized; Extracting a current sentence to be recognized that has the core word from among the multiple sentences to be recognized; Extracting the current associated words and the number of the current associated words in each of the current sentences to be recognized; Calculating the relevance of the current sentence to be recognized according to the preset weight of the current associated word and the current number of associated words in the current sentence to be recognized; Taking the current sentence to be recognized whose relevance is greater than a second threshold as the text area corresponding to the target key content category; In the text data to be recognized, target key content categories corresponding to a plurality of the text regions are marked.

2. The task processing method of smart office according to claim 1, characterized in that: The step of determining target secondary classification information of the text data to be recognized according to the multiple initial secondary classification information corresponding to each of the multiple target word sets comprises: According to the mapping relationship between the first word, the target word set and the initial secondary classification information, the number of initial secondary classification information mapped to each first word is counted; The first word whose amount of initial secondary classification information is less than the first threshold is used as the target first word; Counting multiple initial secondary classification information corresponding to all target first words to form a secondary classification information set; Counting second quantities corresponding to different types of secondary classification information in the secondary classification information set; The secondary classification information corresponding to the maximum second quantity is used as the target secondary classification information.

3. The task processing method of smart office according to claim 1, characterized in that: The step of calculating the relevance of the current sentence to be recognized according to the preset weight of the current associated words and the current number of associated words in the current sentence to be recognized comprises: Adding the preset weights of all current associated words in the current sentence to be recognized to obtain a total weight; Multiplying the total weight by a first adjustment coefficient to obtain a first sub-score; Multiplying the current number of associated words by a second adjustment coefficient to obtain a second sub-score; The first sub-score and the second sub-score are added together to obtain the degree of association.

4. The task processing method of smart office as claimed in claim 1, characterized in that: Before the step of obtaining the preset weight corresponding to the associated word, the method further includes: Obtain multiple sample sentences with core words; Performing word segmentation processing on the plurality of sample sentences respectively to obtain a plurality of sample words; Among all the sample words, obtain the number of occurrences of different sample words to be counted; Dividing the number of times by the total number of sample words to obtain the word frequency; Counting target sample sentences having the sample words to be counted; Dividing the total number of sample sentences by the target number of sample sentences to obtain a first value; Calculating the logarithm of the first value to obtain a rarity index of the sample word to be counted; The word frequency corresponding to each sample word to be counted is multiplied by the rarity index to obtain the preset weights corresponding to different sample words to be counted.

5. The task processing method of smart office as claimed in claim 4, characterized in that: The step of multiplying the word frequency corresponding to each sample word to be counted by the rarity index to obtain the preset weights corresponding to different sample words to be counted comprises: Multiply the word frequency and rarity index of each sample word to be counted to get the initial weight; When the initial weight of the sample word pair to be counted is greater than the third threshold, the initial weight is used as the preset weight corresponding to the sample word to be counted; When the initial weight of the sample word pair to be counted is not greater than the third threshold, the sample word to be counted is eliminated.

6. A task processing device for smart office, characterized in that: The task processing device of the smart office includes: A first acquisition unit is used to acquire text data to be identified and primary classification information of the text data to be identified; the primary classification information includes research reports, paper journals or contracts; A first extraction unit, used for extracting a current title part and / or a current abstract part in the text data to be recognized; A determination unit is used to perform word segmentation processing on the current title part and / or the current abstract part to obtain multiple first words; obtain all the secondary classification information corresponding to the primary classification information; obtain multiple word sets corresponding to all the secondary classification information; wherein each of the word sets includes preset sample words obtained after word segmentation processing of a single sample title part or a single sample abstract part; for each first word, match a target word set having the first word; determine the target secondary classification information of the text data to be identified according to multiple initial secondary classification information corresponding to each of the multiple target word sets; the target secondary classification information is the field category corresponding to each of the primary classification information; The second acquisition unit is used to acquire a target key content category corresponding to the target secondary classification information and a keyword corresponding to the target key content category; the target key content category is a key content selected by a user from a plurality of preset key content categories; the keyword includes a core word and an associated word; The second extraction unit is used to obtain the preset weight corresponding to the associated word; process the text data to be recognized by sentence segmentation to obtain multiple sentences to be recognized; extract the current sentence to be recognized with the core word from the multiple sentences to be recognized; extract the current associated words and the number of current associated words in each of the current sentences to be recognized; calculate the relevance of the current sentence to be recognized according to the preset weight of the current associated words and the number of current associated words in the current sentence to be recognized; and use the current sentence to be recognized with a relevance greater than a second threshold as the text area corresponding to the target key content category; The marking unit is used to mark the target key content categories corresponding to the multiple text areas in the text data to be recognized.

7. A terminal device, characterized in that: The terminal device includes: a memory, a processor, and a smart office task processing program stored in the memory and executable on the processor, wherein the smart office task processing program is configured to implement the steps in the smart office task processing method as described in any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps in the task processing method for smart office as described in any one of claims 1 to 5 are implemented.

Images