Archive resource intelligent recommendation method and system for engineering design report writing
Through multi-dimensional evaluation of vocabulary importance and intelligent matching of key terms, the problem of inaccurate keyword matching in the writing of engineering design reports is solved, and more accurate and high-quality archival resource recommendations are achieved.
Patent Information
- Application Number
- CN202510020147.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-07
- Publication Date
- 2025-03-07
AI Technical Summary
In the process of compiling engineering design reports, it is difficult to distinguish between demand differences in different backgrounds through simple keyword matching, resulting in insufficient recommendation results and a large number of false positives and missed reports.
Multi-dimensional evaluation of vocabulary importance, including TF-IDF algorithm, location weights, contextual relevance, topic relevance and user behavior data, intelligently match key terms, and generate a list of file resource recommendations through dynamic recommendation mechanisms.
It significantly improves the quality of recommendation information, reduces false alarms and missed reports, makes the recommendation results more accurate and can better meet the specific needs of users.
Smart Images
Figure CN119577065A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of archive recommendation, and in particular to an intelligent recommendation method and system for archive resources used for writing engineering design reports. Background Art
[0002] With the development of modern engineering technology, engineering projects have become more and more complex, and the standards, specifications, and literature required for reference are increasing. The traditional manual method of searching for information is time-consuming and labor-intensive, and due to the frequent update of information, it is difficult to ensure that the information obtained is the latest and most accurate. In addition, in the process of team collaboration in large-scale engineering projects, the efficiency of information sharing between different members is low, which often leads to duplication of work or information lag.
[0003] There are some document management and resource sharing platforms on the market. Although they can solve the problem of information sharing to a certain extent, in terms of precision, the existing systems mainly rely on simple keyword matching for recommendation, which is prone to a large number of false positives and false negatives. Specifically: In the process of writing engineering design reports, designers need to filter out relevant archival resources from a large number of documents. When using simple keyword matching, the system will recommend many irrelevant documents, causing designers to spend a lot of time filtering and excluding irrelevant information; For the same set of keywords, engineering design projects in different contexts may have different requirements, and existing systems are often unable to distinguish these differences, resulting in inaccurate recommendation results. Summary of the invention
[0004] In order to solve the above technical problems, the present invention provides an intelligent recommendation method and system for archival resources for engineering design report writing, so as to realize multi-dimensional evaluation of vocabulary importance, refined calculation of position weights, consideration of user behavior data, introduction of topic relevance, intelligent matching of key terms and dynamic recommendation mechanism, which significantly improves the quality of recommended information.
[0005] The present invention provides an intelligent recommendation method for archive resources for writing engineering design reports, comprising: Step S1: receiving engineering design related information input by a user; Step S2: extracting key elements of engineering design based on the relevant information; Step S21: Calculate the importance of each word in the relevant information using the TF-IDF algorithm: obtain the basic value, position weight, context relevance, and topic relevance of the word in all archive resources, obtain user behavior data, obtain user behavior degree based on the user behavior data, and obtain the importance of the word based on the basic value, position weight, context relevance, topic relevance, and user behavior degree; Step S22: constructing a set of predefined engineering design elements; Step S23: sorting the importance of the words in the relevant information from high to low, and taking the top N words as key terms; Step S24: matching the key terms with a set of predefined engineering design elements, generating a mapping table based on the matching results, and extracting the key elements from the mapping table; Step S3: Determine the corresponding engineering design type according to the extracted key elements; Step S4: According to the engineering design type, retrieve archival resources from the knowledge graph and generate a recommended list of archival materials.
[0006] Furthermore, the position weight of the word in all archive resources is obtained, including: Step Sa1, taking the first character of the title in each archive resource as the starting position; taking the last character of each archive resource as the ending position; determining the character length L of the archive resource according to the starting position and the ending position; defining the fixed weight of the character length of the title as W; Step Sa2, constructing a first position threshold, a second position threshold, a third position threshold and a fourth position threshold at positions of different character lengths from the starting position; Step Sa3: construct a formula group to calculate the position weight based on the fixed weight, the distance between the vocabulary and the initial position, the first position threshold, the second position threshold, the third position threshold and the fourth position threshold.
[0007] Furthermore, the archive resources are divided into main titles, subtitles and document bodies; the main title has a character length of m characters, the subtitle has a character length range of [m+1, f], the first position threshold is constructed at a position f characters long from the starting position, the fourth position threshold is constructed at the last character at the end of the document body, and the second position threshold and the third position threshold are constructed based on the first position threshold and the fourth position threshold.
[0008] Furthermore, the position weight formula group is: ; ; In the formula, represents the position weight of the Sth word in the Zth archive resource, Represents the fixed weight of the character length of the Zth archive resource title. The threshold function representing the distance between the Sth word and the starting position in the Zth archive resource; and Respectively represent the position weight attenuation function of the Z-th archive resource at the first position threshold and the second position threshold. Specifically, and Represent the decay index, m represents the character length of the title, represents the first position threshold, represents the second position threshold, and They represent the position weight rapid decrease function and weight accelerated increase function of the Zth archive resource at the third position threshold and the fourth position threshold, respectively. represents the third position threshold, represents the fourth position threshold, c represents the adjustment coefficient, Represents the distance between the Sth word and the starting position in the Zth archive resource.
[0009] Furthermore, the second position threshold and the third position threshold are constructed according to the first position threshold and the fourth position threshold, including: Step Sb1, presetting the second position threshold to be constructed at the nth character of the document text, and the character length of the second position threshold from the initial position is f+n; presetting the third position threshold to be constructed at the middle character of the document text, and the character length of the third position threshold from the initial position is (Lf) / 2; Step Sb2, based on the first position threshold, construct the character length range [1, f] between the main title and the subtitle; based on the second position threshold, construct the character length range [m+1, f+n] between the subtitle and the nth character of the document text; based on the third position threshold, construct the character length range [f+n+1, (Lf) / 2] between the n+1th character of the document text and the middle character of the document text; based on the fourth position threshold, construct the character length range [{(Lf) / 2}+1, L] between the middle character of the document text and the last character at the end of the document body; Step Sb3, constructing a weighted piecewise function according to the character length range and the distance between the vocabulary and the starting position in step Sb2; Step Sb4, taking the derivative of the weighted piecewise function to determine the character length range of the distance between the vocabulary and the initial position when the derivative is zero; according to the character length range of the distance between the vocabulary and the initial position when the derivative is zero, calculating the maximum and minimum values of the weighted piecewise function; Step Sb5, determining the position of the vocabulary when the weighted piecewise function reaches the minimum value according to the minimum value and the construction position of the second position threshold, and reconstructing the third position threshold at the position of the vocabulary when the weighted piecewise function reaches the minimum value; Step Sb6, determining the position of the vocabulary when the weight piecewise function reaches the maximum value according to the maximum value and the construction position of the third position threshold, and reconstructing the second position threshold at the position of the vocabulary when the weight piecewise function reaches the maximum value.
[0010] Furthermore, the acquisition of the topic relevance includes: acquiring the topic distribution of words and the topic distribution of archive resources through a topic model, and calculating the topic relevance according to the topic distribution of words, the topic distribution of archive resources, and the position weight of words.
[0011] Furthermore, the calculation formula of the topic relevance is: ; In the formula, represents the topic relevance of the Sth word in the Zth archive resource, represents the position weight of the Sth word in the Zth archive resource, represents the topic distribution of the Sth word in the Zth archive resource, Represents the topic distribution of the Zth archive resource.
[0012] Furthermore, the calculation formula of the vocabulary importance is: ; In the formula, represents the importance of the Sth word in the Zth archive resource, represents the Sth lexical base value in the Zth archive resource, represents the topic relevance of the Sth word in the Zth archive resource, represents the contextual relevance of the Sth word in the Zth archive resource, represents the user behavior degree of the Sth word in the Zth archive resource, Represents the position weight of the Sth word in the Zth archive resource.
[0013] The present invention also provides an intelligent recommendation system for archival resources used for writing engineering design reports, comprising the following modules: Related information receiving module: used to receive engineering design related information input by the user; Key element extraction module: connected to the relevant information receiving module, used to extract key elements of engineering design based on the relevant information; Engineering design type determination module: connected with the key element extraction module, used to determine the corresponding engineering design type according to the extracted key elements; Archive recommendation module: connected to the engineering design type determination module, it is used to retrieve archive resources from the knowledge graph and generate a recommended list of archive materials based on the engineering design type.
[0014] The embodiments of the present invention have the following technical effects: Through the technical means eliminated by this application, the present invention effectively solves the problem in the prior art that simple keyword matching cannot distinguish the differences in needs under different backgrounds, shortens the time for file screening, and improves the accuracy of recommended information. Specifically, the present invention significantly improves the quality of recommended information, reduces false positives and false negatives, and makes the recommendation results more accurate and better meets the specific needs of users by evaluating the importance of vocabulary in multiple dimensions, calculating position weights in a refined manner, considering user behavior data, introducing topic relevance, intelligently matching key terms, and dynamic recommendation mechanisms.
[0015] The present invention calculates the importance of words through the TF-IDF algorithm, which not only considers the basic value of words, but also combines multiple dimensions such as position weight, context relevance, topic relevance and user behavior. This enables the system to more comprehensively understand the meaning of words in different contexts, thereby improving the accuracy of recommendations.
[0016] The position weight of a word is constructed by setting four position thresholds to ensure that the position information of the word in the archival resources is accurately reflected. The precise calculation of the position weight helps to distinguish the importance of the same word in different positions, thereby reducing false positives and false negatives.
[0017] When calculating the importance of words, the present application takes into account the user's behavioral data, so that the recommendation results are closer to the user's actual needs and increase the level of personalization of the recommendations. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0019] Figure 1 It is a flow chart of an intelligent recommendation method for archival resources for writing engineering design reports provided by an embodiment of the present invention; Figure 2 It is a structural schematic diagram of an intelligent recommendation method for archival resources for writing engineering design reports provided by an embodiment of the present invention; Figure 3 It is a structural schematic diagram of an electronic device provided by the implementation of the present invention. DETAILED DESCRIPTION
[0020] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work belong to the scope of protection of the present invention.
[0021] Figure 1 Flowchart of an intelligent recommendation method for archival resources for engineering design report writing provided by an embodiment of the present invention. Figure 1 , specifically including: Step S1: receiving engineering design related information input by a user.
[0022] For the person in charge or related personnel responsible for engineering design, it is necessary to log in to the engineering design software platform and input various engineering design related information into the engineering design software platform to query the archival resources used for writing engineering design reports.
[0023] Engineering design related information includes but is not limited to: Basic information: project name, project number, project location, project type, design stage, etc.; Time information: project start time, project end time, etc.; Personnel information: responsible manager, responsible team members, etc.; Documents: Design drawings, design reports, etc. that can be uploaded; Keywords related to the current engineering design, which will be used for subsequent information retrieval.
[0024] The purpose of the user entering the above information is to find archival resources for writing engineering design reports. This information will help the system understand the user's needs more accurately and recommend appropriate document resources accordingly.
[0025] Step S2: extracting key elements of engineering design based on the relevant information.
[0026] Step S21: Use the TF-IDF algorithm to calculate the importance of each word in the relevant information: obtain the basic value, position weight, context relevance, and topic relevance of the word in all archival resources, obtain user behavior data, and obtain the user behavior degree based on the user behavior data, and obtain the importance of the word based on the basic value, position weight, context relevance, topic relevance, and user behavior degree.
[0027] Get the position weight of the word in all archive resources, including: Step Sa1, take the first character of the main title in each file resource as the starting position; take the last character of each file resource as the ending position; determine the character length L of the file resource according to the starting position and the ending position; define the fixed weight of the character length where the main title is located as W.
[0028] Exemplarily, if the main title of the document is "Engineering Design Summary", then the character position where the character "工" is located is the starting position. The last line of the document is "October 22, 2024", then the character position where the character "日" is located is the ending position. If the starting position is the 1st character and the ending position is the 1000th character, then the character length L = 1000. In this embodiment, it is preferably set that the weight of each character within the character length of the main title of all file resources is a fixed weight and is the maximum weight in the entire file resource. This embodiment believes that the main title has the highest importance in the entire file resource. The fixed weight is set as W.
[0029] Step Sa2, construct a first position threshold, a second position threshold, a third position threshold, and a fourth position threshold at positions with different character lengths from the starting position.
[0030] Divide the file resource into a main title, a subtitle, and the document body; the character length of the main title is m characters, the character length range of the subtitle is [m + 1, f], the first position threshold is constructed at the position with a character length of f from the starting position, the fourth position threshold is constructed at the last character at the end of the document body, and the second position threshold and the third position threshold are constructed according to the first position threshold and the fourth position threshold.
[0031] The prior art believes that the main title has the highest weight, the subtitle is second, and the beginning, middle, and end of the document body are inferior to the subtitle. Therefore, in this embodiment, the main title is used as the reference position, and a first position threshold, a second position threshold, a third position threshold, and a fourth position threshold are set at different positions of the file resource, and the position weight is analyzed by the vocabulary being at different distances from the initial position. According to the fact that the subtitle of the file resource in the prior art is more important than the document body, in this embodiment, the first position threshold is set at the last character of the subtitle, and the fourth position threshold is set at the last character at the end of the document body.
[0032] The beginning of a document contains a variety of information, such as an abstract, a table of contents, a copyright statement, etc. The importance of this information is not necessarily lower than that of the main text, so the position of the beginning of the document is usually not a fixed position, but changes with the content of the document. The content distribution of the document is also uneven. Some documents may have a large number of tables or charts in the middle, while other documents have more text descriptions in the middle. Therefore, the importance of the middle position varies from document to document. In this regard, this embodiment proposes a construction process of the first position threshold and the fourth position threshold. Specifically: the second position threshold and the third position threshold are constructed based on the first position threshold and the fourth position threshold, including: Step Sb1, preset the second position threshold to be constructed at the nth character of the document text, and the character length of the second position threshold from the initial position is f+n; preset the third position threshold to be constructed at the middle character of the document text, and the character length of the third position threshold from the initial position is (Lf) / 2.
[0033] Step Sb2, based on the first position threshold, construct the character length range A1=[1, f] between the main title and the subtitle; based on the second position threshold, construct the character length range A2=[m+1, f+n] between the subtitle and the nth character of the document text; based on the third position threshold, construct the character length range A3=[f+n+1, (Lf) / 2] between the n+1th character of the document text and the middle character of the document text; based on the fourth position threshold, construct the character length range A4=[{(Lf) / 2}+1, L] between the middle character of the document text and the last character at the end of the document body.
[0034] Step Sb3, constructing a weighted piecewise function according to the character length range and the distance between the vocabulary and the starting position in step Sb2.
[0035] This embodiment constructs a weighted piecewise function based on the distance between the vocabulary and the initial position and the character length range: ; represents the weighted piecewise function constructed based on the distance between the vocabulary and the initial position, d represents the distance between the vocabulary and the initial position, represents the coefficients of the first piecewise function, represents the coefficients of the second piecewise function, represents the coefficients of the third piecewise function, Represents the coefficients of the fourth piecewise function. , are positive numbers respectively.
[0036] Step Sb4, derive the weighted piecewise function to determine the character length range of the distance between the vocabulary and the initial position when the derivative is zero; calculate the maximum and minimum values of the weighted piecewise function based on the character length range of the distance between the vocabulary and the initial position when the derivative is zero.
[0037] In order to determine the minimum position of the weight piecewise function, the derivative of the weight change function is calculated, and the point where the derivative is zero is found. There is no minimum and maximum value for the first piecewise function and the second piecewise function.
[0038] exist , the derivative is zero: =0; ; This means that the vocabulary is in Within the character length range, weight time segment function can reach a minimum value.
[0039] The minimum value is: .
[0040] exist , the derivative is zero: =0; ; This means that the vocabulary is in Within the character length range, weight time segment function Can reach a maximum value. The maximum value is: .
[0041] Step Sb5, determining the position of the vocabulary when the weight piecewise function reaches the minimum value according to the minimum value and the construction position of the second position threshold, and reconstructing the third position threshold at the position of the vocabulary when the weight piecewise function reaches the minimum value.
[0042] Step Sb6, determining the position of the vocabulary when the weight piecewise function reaches the maximum value according to the maximum value and the construction position of the third position threshold, and reconstructing the second position threshold at the position of the vocabulary when the weight piecewise function reaches the maximum value.
[0043] According to the above analysis, the minimum value occurs in the range of A2, and the maximum value occurs in the range of A3. The second position threshold and the third position threshold are reset according to these positions. For example, the minimum value is in the range of A2=[m+1, f+n], and the preset minimum value is in the range of d xAt the position, the range length of A2 is (f+n)-(m+1), and the new position threshold is calculated by the character length L of the archive resource, the minimum preset position, the lower limit of the character length range to which the current threshold belongs, and the length.
[0044] Exemplarily, the calculation formula for the obtained second position threshold is: .
[0045] Step Sa3: construct a formula group to calculate the position weight based on the fixed weight, the distance between the vocabulary and the initial position, the first position threshold, the second position threshold, the third position threshold and the fourth position threshold.
[0046] ; ; In the formula, represents the position weight of the Sth word in the Zth archive resource, Represents the fixed weight of the character length of the Zth archive resource title. The threshold function representing the distance between the Sth word and the starting position in the Zth archive resource; and Respectively represent the position weight attenuation function of the Z-th archive resource at the first position threshold and the second position threshold. Specifically, and Represent the decay exponent, , m represents the character length of the title, represents the first position threshold, represents the second position threshold, and They represent the position weight rapid decrease function and weight accelerated increase function of the Zth archive resource at the third position threshold and the fourth position threshold, respectively. represents the third position threshold, represents the fourth position threshold, c represents the adjustment coefficient, Represents the distance between the Sth word and the starting position in the Zth archive resource.
[0047] Since the importance of the subheading is second only to the main headline, and the weight of the main headline is consistent throughout the entire length, because the main headline contains an overview of the core theme or main content of the document, it is very important throughout the entire length. Although the importance of the main headline is very high, the subheading that follows it is also very important. Therefore, the weight gradually decreases from the end of the main headline to the end of the subheading, but the rate of decrease is slow to reflect that this area is still of high importance. The part after the subheading contains some transitional information, such as an explanation or a brief introduction to the subheading. The importance of this information is still high, but not as important as the subheading itself. Therefore, the weight continues to decrease slowly in this area. Therefore, when the vocabulary is at the first position threshold, between the character after the first position threshold and the character before the second position threshold, the present embodiment sets two attenuation functions respectively, the difference being that: the attenuation coefficient of the attenuation function when the vocabulary is at the first position threshold The attenuation coefficient of the attenuation function between the character after the first position threshold and the character before the second position threshold .
[0048] The document content within the character length between the second position threshold and the character before the third position threshold belongs to a part of the document body. Although this part of the content is important, its importance gradually decreases relative to the title part. Therefore, the weight drops rapidly in this area, reflecting the decrease in the importance of the body relative to the title. Therefore, when the vocabulary is at the position of the character before the second position threshold to the third position threshold, the present embodiment sets a rapid decline function. The part after the third position threshold is usually close to the end of the document. This part of the content contains information such as a summary or conclusion, and its importance is increased. Therefore, the weight rises rapidly in this area, reflecting the increase in the importance of the ending part relative to the middle body part. The present embodiment sets an accelerated increasing function within this range.
[0049] The acquisition of the topic relevance includes: acquiring the topic distribution of words and the topic distribution of archive resources through a topic model, and calculating the topic relevance according to the topic distribution of words, the topic distribution of archive resources, and the position weight of words.
[0050] ; In the formula, represents the topic relevance of the Sth word in the Zth archive resource, represents the position weight of the Sth word in the Zth archive resource, represents the topic distribution of the Sth word in the Zth archive resource, Represents the topic distribution of the Zth archive resource.
[0051] The context relevance is obtained by extracting context words around the current word, and calculating the word vectors of the current word and the context words using a pre-trained word vector model. The average value of the context word vector and the current word vector is calculated, and the cosine similarity between the current word vector, the context word vector, and the average value of the context word vector and the current word vector is used as the context relevance of this embodiment.
[0052] To obtain the user behavior degree, first collect user behavior data (user behavior data includes user browsing behavior, download records, feedback information, etc.), standardize the user behavior data, and filter indicators from the standardized user behavior data. The indicators may include: number of clicks, dwell time, number of return visits, number of downloads, number of feedbacks, and feedback scores. Construct a matrix with the number of clicks, dwell time, number of return visits, number of downloads, number of feedbacks, and feedback scores, calculate the covariance matrix of the matrix, and calculate the eigenvalues and eigenvectors of the covariance matrix. Sort the eigenvalues from high to low, take the eigenvectors of the first 1 / 4 eigenvalue sets as the main components, and take the eigenvectors of the main components as the weight vectors. Multiply the weights in the weight vectors and the corresponding indicators to obtain the user behavior degree.
[0053] Among them, the term frequency TF obtained by the TF-IDF algorithm is multiplied by the inverse document frequency IDF to obtain the basic value.
[0054] The calculation formula of the word importance is: ; In the formula, represents the importance of the Sth word in the Zth archive resource, represents the Sth lexical base value in the Zth archive resource, represents the topic relevance of the Sth word in the Zth archive resource, represents the contextual relevance of the Sth word in the Zth archive resource, represents the user behavior degree of the Sth word in the Zth archive resource, Represents the position weight of the Sth word in the Zth archive resource.
[0055] Step S22: construct a set of predefined engineering design elements.
[0056] This embodiment takes a construction project as an example. First, the basic elements of engineering design are determined, such as building type (residential, commercial, industrial), building materials (concrete, steel, wood), structure type (frame structure, shear wall structure), functional area (bedroom, kitchen, office), building height, number of floors, safety standards, environmental protection standards, etc. Each basic element is further subdivided. For example, for building materials, it can be subdivided into reinforced concrete, prefabricated components, glass curtain walls, etc. Finally, a detailed set of engineering design elements is constructed.
[0057] Step S23: sort the importance of the words in the relevant information from high to low, and use the first N words as key terms.
[0058] Extract all mentioned terms from relevant documents, reports or other information sources and rank them according to their importance in the documents. Then select the most important N terms as key terms.
[0059] Step S24: matching the key terms with a set of predefined engineering design elements, generating a mapping table based on the matching results, and extracting key elements from the mapping table.
[0060] The key terms obtained in step S23 are matched with the predefined engineering design element set created in step S22, for example, "reinforced concrete" is matched with "building materials", and "seismic level" is matched with "safety standards". A mapping table is constructed to show the matching relationship.
[0061] Table 1 Mapping table of key terms and predefined engineering design element sets
[0062] Extract key elements from the mapping table. For example, the following key elements can be extracted from the mapping table: Building materials: reinforced concrete; Safety standards: earthquake resistance level; Environmental protection standards: energy-saving standards; Structural type: shear wall structure; Number of floors: three-story building; Building height: multi-story.
[0063] This embodiment can quickly identify and focus on key elements in engineering design based on the importance analysis results of archival resources, improve the efficiency of information processing, and ensure that the most important aspects are taken into account during the engineering design process.
[0064] Step S3: Determine the corresponding engineering design type based on the extracted key elements.
[0065] For example, based on the above key elements, it can be seen that the building is a multi-story residential or commercial building, using reinforced concrete as the main building material, and has high earthquake resistance requirements and energy-saving standards. It can be preliminarily determined that the project design type is "multi-story reinforced concrete residential / commercial building" Step S4: According to the engineering design type, retrieve archival resources from the knowledge graph and generate a recommended list of archival materials.
[0066] The knowledge graph contains archival resources of various types of engineering design, including but not limited to construction drawings, material lists, safety specifications, construction guidelines, past project cases, etc. According to the determined engineering design type, the relevant archival resources in the knowledge graph are queried, and a recommended list of archival materials for the engineering design type is generated based on the query results.
[0067] The recommended list of archival materials includes, but is not limited to: construction drawings, material lists, safety specifications, construction guidelines, past project cases, project archives, etc.
[0068] like Figure 2 As shown, an intelligent recommendation system for archival resources for writing engineering design reports includes the following modules: Related information receiving module: used to receive engineering design related information input by the user; Key element extraction module: connected to the relevant information receiving module, used to extract key elements of engineering design based on the relevant information; Engineering design type determination module: connected with the key element extraction module, used to determine the corresponding engineering design type according to the extracted key elements; Archive recommendation module: connected to the engineering design type determination module, it is used to retrieve archive resources from the knowledge graph and generate a recommended list of archive materials based on the engineering design type.
[0069] Figure 3 Schematic diagram of the structure of an electronic device provided by the present invention. Figure 3 As shown, the electronic device 500 includes one or more processors 501 and a memory 502 .
[0070] The processor 501 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 500 to perform desired functions.
[0071] The memory 502 may include one or more computer program products, and the computer program product may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory (cache), etc. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 501 may run the program instructions to implement an intelligent recommendation method for archival resources for writing engineering design reports and / or other desired functions of any embodiment of the present application described above. Various contents such as initial external parameters, thresholds, etc. may also be stored in the computer-readable storage medium.
[0072] In one example, the electronic device 500 may further include: an input device 503 and an output device 504, which are interconnected via a bus system and / or other forms of connection mechanisms (not shown). The input device 503 may include, for example, a keyboard, a mouse, etc. The output device 504 may output various information to the outside, including early warning prompt information, braking force, etc. The output device 504 may include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto, etc.
[0073] Of course, to simplify, Figure 3 Only some of the components related to the present application in the electronic device 500 are shown, and components such as a bus, an input / output interface, etc. are omitted. In addition, according to specific application situations, the electronic device 500 may also include any other appropriate components.
[0074] In addition to the above-mentioned methods and devices, an embodiment of the present application may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to execute the steps of a method for intelligently recommending archival resources for writing engineering design reports provided in any embodiment of the present application.
[0075] The computer program product may be written in any combination of one or more programming languages to write program codes for performing the operations of the embodiments of the present application, including object-oriented programming languages, such as Java, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as an independent software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.
[0076] In addition, an embodiment of the present application may also be a computer-readable storage medium on which computer program instructions are stored. When the computer program instructions are executed by a processor, the processor executes the steps of a method for intelligently recommending archival resources for writing engineering design reports provided in any embodiment of the present application.
[0077] The computer readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can include, for example, but is not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.
[0078] It should be noted that the terms used in the present invention are only for describing specific embodiments, rather than limiting the scope of the present application. As shown in the present specification, unless the context clearly indicates an exception, the words "one", "a", "a kind of" and / or "the" do not specifically refer to the singular, but may also include the plural. The terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method or device. In the absence of more restrictions, the elements defined by the sentence "include one..." do not exclude the presence of other identical elements in the process, method or device including the elements.
[0079] It should also be noted that the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. Unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", etc. should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be a connection between the two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0080] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the technical solutions of the embodiments of the present invention.
Claims
1. An intelligent recommendation method for archival resources for writing engineering design reports, characterized in that: include: Step S1: receiving engineering design related information input by a user; Step S2: extracting key elements of engineering design based on the relevant information; Step S21: Calculate the importance of each word in the relevant information using the TF-IDF algorithm: obtain the basic value, position weight, context relevance, and topic relevance of the word in all archive resources, obtain user behavior data, obtain user behavior degree based on the user behavior data, and obtain the importance of the word based on the basic value, position weight, context relevance, topic relevance, and user behavior degree; Step S22: constructing a set of predefined engineering design elements; Step S23: sorting the importance of the words in the relevant information from high to low, and taking the top N words as key terms; Step S24: matching the key terms with a set of predefined engineering design elements, generating a mapping table based on the matching results, and extracting the key elements from the mapping table; Step S3: Determine the corresponding engineering design type according to the extracted key elements; Step S4: According to the engineering design type, retrieve archival resources from the knowledge graph and generate a recommended list of archival materials.
2. The method for intelligently recommending archive resources for writing engineering design reports according to claim 1, characterized in that: Get the position weight of the word in all archive resources, including: Step Sa1, taking the first character of the title in each archive resource as the starting position; taking the last character of each archive resource as the ending position; determining the character length L of the archive resource according to the starting position and the ending position; defining the fixed weight of the character length of the title as W; Step Sa2, constructing a first position threshold, a second position threshold, a third position threshold and a fourth position threshold at positions of different character lengths from the starting position; Step Sa3: construct a formula group to calculate the position weight based on the fixed weight, the distance between the vocabulary and the initial position, the first position threshold, the second position threshold, the third position threshold and the fourth position threshold.
3. The intelligent recommendation method for archival resources for engineering design report writing according to claim 2 is characterized in that: The archive resources are divided into a main title, a subtitle and a document body; the character length of the main title is m characters, the character length range of the subtitle is [m+1, f], the first position threshold is constructed at a position f characters away from the starting position, the fourth position threshold is constructed at the last character at the end of the document body, and the second position threshold and the third position threshold are constructed based on the first position threshold and the fourth position threshold.
4. The method for intelligently recommending archive resources for writing engineering design reports according to claim 2, characterized in that: The formula group of the position weight is: ； ； In the formula, represents the position weight of the Sth word in the Zth archive resource, Represents the fixed weight of the character length of the Zth archive resource title. The threshold function representing the distance between the Sth word and the starting position in the Zth archive resource; and Respectively represent the position weight attenuation function of the Z-th archive resource at the first position threshold and the second position threshold. Specifically, and Represent the decay index, m represents the character length of the title, represents the first position threshold, represents the second position threshold, and They represent the position weight rapid decrease function and weight accelerated increase function of the Zth archive resource at the third position threshold and the fourth position threshold, respectively. represents the third position threshold, represents the fourth position threshold, c represents the adjustment coefficient, Represents the distance between the Sth word and the starting position in the Zth archive resource.
5. The method for intelligently recommending archive resources for writing engineering design reports according to claim 3, characterized in that: The second position threshold and the third position threshold are constructed according to the first position threshold and the fourth position threshold, including: Step Sb1, presetting the second position threshold to be constructed at the nth character of the document text, and the character length of the second position threshold from the initial position is f+n; presetting the third position threshold to be constructed at the middle character of the document text, and the character length of the third position threshold from the initial position is (Lf) / 2; Step Sb2, based on the first position threshold, construct the character length range [1, f] between the main title and the subtitle; based on the second position threshold, construct the character length range [m+1, f+n] between the subtitle and the nth character of the document text; based on the third position threshold, construct the character length range [f+n+1, (Lf) / 2] between the n+1th character of the document text and the middle character of the document text; based on the fourth position threshold, construct the character length range [{(Lf) / 2}+1, L] between the middle character of the document text and the last character at the end of the document body; Step Sb3, constructing a weighted piecewise function according to the character length range and the distance between the vocabulary and the starting position in step Sb2; Step Sb4, taking the derivative of the weighted piecewise function to determine the character length range of the distance between the vocabulary and the initial position when the derivative is zero; according to the character length range of the distance between the vocabulary and the initial position when the derivative is zero, calculating the maximum and minimum values of the weighted piecewise function; Step Sb5, determining the position of the vocabulary when the weighted piecewise function reaches the minimum value according to the minimum value and the construction position of the second position threshold, and reconstructing the third position threshold at the position of the vocabulary when the weighted piecewise function reaches the minimum value; Step Sb6, determining the position of the vocabulary when the weight piecewise function reaches the maximum value according to the maximum value and the construction position of the third position threshold, and reconstructing the second position threshold at the position of the vocabulary when the weight piecewise function reaches the maximum value.
6. The method for intelligently recommending archive resources for writing engineering design reports according to claim 1, characterized in that: The acquisition of the topic relevance includes: acquiring the topic distribution of words and the topic distribution of archive resources through a topic model, and calculating the topic relevance according to the topic distribution of words, the topic distribution of archive resources, and the position weight of words.
7. The method for intelligently recommending archive resources for writing engineering design reports according to claim 6, characterized in that: The calculation formula of the topic relevance is: ； In the formula, represents the topic relevance of the Sth word in the Zth archive resource, represents the position weight of the Sth word in the Zth archive resource, represents the topic distribution of the Sth word in the Zth archive resource, Represents the topic distribution of the Zth archive resource.
8. The method for intelligently recommending archive resources for writing engineering design reports according to claim 5, characterized in that: The calculation formula of the word importance is: ； In the formula, represents the importance of the Sth word in the Zth archive resource, represents the Sth lexical base value in the Zth archive resource, represents the topic relevance of the Sth word in the Zth archive resource, represents the contextual relevance of the Sth word in the Zth archive resource, represents the user behavior degree of the Sth word in the Zth archive resource, Represents the position weight of the Sth word in the Zth archive resource.
9. An intelligent recommendation system for archival resources used for writing engineering design reports, used to execute an intelligent recommendation method for archival resources used for writing engineering design reports as described in any one of claims 1 to 8, characterized in that: Includes the following modules: Related information receiving module: used to receive engineering design related information input by the user; Key element extraction module: connected to the relevant information receiving module, used to extract key elements of engineering design based on the relevant information; Engineering design type determination module: connected with the key element extraction module, used to determine the corresponding engineering design type according to the extracted key elements; Archive recommendation module: connected to the engineering design type determination module, it is used to retrieve archive resources from the knowledge graph and generate a recommended list of archive materials based on the engineering design type.