House building project engineering archive data management system and method thereof

Abstract

The invention discloses a house building project engineering archive data management system and method, and the method comprises the steps: carrying out the segment granularity theme mining of the house building project engineering archive data uploaded by a user through employing a natural language processing technology based on deep learning, so as to extract the main content of the archive data; semantic encoding is carried out on each obtained paragraph theme, semantic feature representation of each paragraph theme is captured, and semantic enhancement processing is further carried out based on context semantic association structures among the paragraph themes, so that deep semantic understanding of global context information of the housing construction project archive data is realized; and on the basis, a specific type label of the archive data is intelligently judged and allocated, so that automatic and intelligent archiving of the housing construction project archive data is realized. In this way, the efficiency and accuracy of housing construction project engineering archive data management can be effectively improved, manual intervention is reduced, and the risk of misjudgment and missed judgment is reduced.

Description

Technical Field

[0001] The present application relates to the field of intelligent management, and more specifically, to a building construction project engineering archive data management system and method thereof. Background Art

[0002] With the rapid development of the construction industry, the demand for the management and archiving of construction project archives is increasing. These archives include but are not limited to design drawings, construction records, acceptance reports, contract documents, etc., which are of various types and huge in quantity. Traditional archive management methods mainly rely on manual operation, which is prone to misjudgment and omission due to the subjectivity and fatigue of manual operation, resulting in incorrect classification of archives, affecting subsequent query and use.

[0003] In order to overcome the above problems, the development of modern information technology has provided a new solution for the management of construction project archives. Although the existing digital archive management system has improved management efficiency to a certain extent, it is difficult to accurately understand the deep semantics of documents when processing complex documents because the existing system relies on preset rules and simple keyword matching to classify archives, resulting in inaccurate classification, which directly affects the efficiency and accuracy of archive management.

[0004] Therefore, an optimized building construction project engineering archive data management system and method thereof are expected. Summary of the invention

[0005] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides a housing construction project engineering archive data management system and method thereof, which performs segment granularity topic mining on the housing construction project engineering archive data uploaded by users by utilizing natural language processing technology based on deep learning to extract the main content of the archive data, and semantically encodes the obtained each paragraph theme to capture the semantic feature representation of each paragraph theme, and further, based on the contextual semantic association structure between each paragraph theme, performs semantic enhancement processing on it to achieve a deep semantic understanding of the global context information of the housing construction project engineering archive data, and based on this, intelligently judges and assigns specific type labels of the archive data, thereby realizing the automatic intelligent archiving of the housing construction project engineering archive data. In this way, the efficiency and accuracy of the management of housing construction project engineering archive data can be effectively improved, human intervention can be reduced, and the risk of misjudgment and missed judgment can be reduced.

[0006] According to one aspect of the present application, a method for managing engineering archives of a building construction project is provided, which includes:

[0007] Obtain the engineering archives of housing construction projects uploaded by users;

[0008] Performing segment-based content theme mining on the engineering archives of the housing construction project to obtain a sequence of archival archive segment content theme descriptions;

[0009] Semantically encoding each archival data segment content subject description in the sequence of archival data segment content subject descriptions to obtain a sequence of archival data segment content subject semantic encoding vectors;

[0010] Performing semantic feature enhancement on the sequence of the archival material segment content subject semantic coding vectors to obtain a sequence of archival material segment content subject semantic enhancement coding vectors;

[0011] Cascading the sequence of the archival material segment content subject semantics reinforcement coding vectors to obtain the archival material global content subject semantics cascade coding vector;

[0012] Based on the global content theme semantic cascade coding vector of the archival material, an intelligent archiving result is determined, and based on the intelligent archiving result, the engineering archival material of the housing construction project is archived.

[0013] According to another aspect of the present application, a building construction project engineering archive data management system is provided, which includes:

[0014] The housing construction project engineering archive data acquisition module is used to acquire the housing construction project engineering archive data uploaded by the user;

[0015] An archive data subject mining module is used to perform content subject mining based on segment granularity on the engineering archive data of the housing construction project to obtain a sequence of archival data segment content subject descriptions;

[0016] An archive data segment content theme semantic coding module, used for semantically coding each archive data segment content theme description in the sequence of archive data segment content theme descriptions to obtain a sequence of archive data segment content theme semantic coding vectors;

[0017] The archival material segment content theme semantic enhancement module is used to enhance the semantic features of the sequence of archival material segment content theme semantic coding vectors to obtain a sequence of archival material segment content theme semantic enhancement coding vectors;

[0018] A cascading module, used for cascading the sequence of the archival material segment content subject semantics reinforcement coding vectors to obtain the archival material global content subject semantics cascade coding vector;

[0019] The intelligent archiving module is used to determine the intelligent archiving result based on the global content theme semantic cascade coding vector of the archival material, and archive the engineering archival material of the housing construction project based on the intelligent archiving result.

[0020] Compared with the prior art, the present application provides a housing construction project engineering archive data management system and method thereof, which uses natural language processing technology based on deep learning to perform segment-granular topic mining on the housing construction project engineering archive data uploaded by users to extract the main content of the archive data, and by semantically encoding the obtained each paragraph theme, capturing the semantic feature representation of each paragraph theme, and further, based on the contextual semantic association structure between each paragraph theme, it performs semantic enhancement processing on it to achieve a deep-level semantic understanding of the global context information of the housing construction project engineering archive data, and based on this, intelligently judge and assign specific type labels of the archive data, thereby realizing the automated intelligent archiving of the housing construction project engineering archive data. In this way, the efficiency and accuracy of the management of housing construction project engineering archive data can be effectively improved, human intervention can be reduced, and the risk of misjudgment and missed judgment can be reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0022] Figure 1 A flowchart of a method for managing construction project archives according to an embodiment of the present application;

[0023] Figure 2 A data flow diagram of a method for managing construction project archives according to an embodiment of the present application;

[0024] Figure 3 It is a flowchart of sub-step S4 of the method for managing engineering archives of a building construction project according to an embodiment of the present application;

[0025] Figure 4 It is a block diagram of a building construction project engineering archive data management system according to an embodiment of the present application. DETAILED DESCRIPTION

[0026] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0027] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0028] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0029] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. Meanwhile, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0030] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0031] Although the existing digital archive management system has improved management efficiency to a certain extent, it is difficult to accurately understand the deep semantics of documents when processing complex documents because the existing system relies on preset rules and simple keyword matching to classify archives, resulting in inaccurate classification, which directly affects the efficiency and accuracy of archive management. Therefore, an optimized house construction project archive management system and method are expected.

[0032] Based on this, in the technical solution of this application, a method for managing engineering archives of building construction projects is proposed. Figure 1 The present invention is a flowchart of a method for managing engineering archives of a building construction project according to an embodiment of the present application. Figure 2 Schematic diagram of data flow of the method for managing archives of building construction projects according to an embodiment of the present application. Figure 1 and Figure 2As shown, according to the embodiment of the present application, the method for managing engineering archive data of a building construction project includes the following steps: S1, obtaining engineering archive data of a building construction project uploaded by a user; S2, performing segment-based content theme mining on the engineering archive data of the building construction project to obtain a sequence of archival data segment content theme descriptions; S3, performing semantic encoding on each archival data segment content theme description in the sequence of archival data segment content theme descriptions to obtain a sequence of archival data segment content theme semantic encoding vectors; S4, performing semantic feature enhancement on the sequence of archival data segment content theme semantic encoding vectors to obtain a sequence of archival data segment content theme semantic enhancement encoding vectors; S5, cascading the sequence of archival data segment content theme semantic enhancement encoding vectors to obtain an archival data global content theme semantic cascade encoding vector; S6, determining an intelligent archiving result based on the archival data global content theme semantic cascade encoding vector, and archiving the building construction project engineering archive data based on the intelligent archiving result.

[0033] In particular, S1 obtains the engineering archives of the housing construction project uploaded by the user. Among them, the engineering archives of the housing construction project refer to various documents, drawings, reports, contracts, photos, videos and other materials with preservation value formed in the whole process of housing construction project from project establishment, design, construction, completion to delivery and use. These materials comprehensively record all aspects of the project from the beginning to the end, and are an important basis for project management and subsequent maintenance. Therefore, archiving the engineering archives of the housing construction project is an important means to ensure the legality and compliance of the project, improve quality, optimize management, prevent risks, inherit history and improve economic benefits.

[0034] In particular, the S2 is to perform content theme mining based on segment granularity on the engineering archives of the housing construction project to obtain a sequence of archival material segment content theme descriptions. In an embodiment of the present application, first, the engineering archives of the housing construction project are subjected to text recognition to obtain the text portion of the engineering archives of the housing construction project. Here, considering that the sources of the engineering archives of the housing construction project are diverse, they may include paper documents, scanned copies, handwritten notes, electronic documents and other formats, and cannot be directly processed. Therefore, in the technical solution of the present application, the engineering archives of the housing construction project are subjected to text recognition to obtain the text portion of the engineering archives of the housing construction project. Among them, the text recognition technology can process these different formats of data, and uniformly convert the data of various formats into a standard text format to ensure that all information can be effectively processed by the system. In one example, OCR technology can be used to perform text recognition on the engineering archives of the housing construction project, which obtains the text information in the archives through a scanner or a camera, and then converts the text in the image into a computer-readable text format through image processing and pattern recognition technology. Next, the text part of the engineering archives of the housing construction project is segmented to obtain a sequence of segment descriptions of the engineering archives of the housing construction project. In the technical solution of the present application, considering that each paragraph of the text part of the engineering archives of the housing construction project may be developed around different central ideas or themes, it is difficult to capture the local semantic information of each independent part by directly analyzing the entire text part. Therefore, in the technical solution of the present application, the text part of the engineering archives of the housing construction project is segmented to obtain a sequence of segment descriptions of the engineering archives of the housing construction project. By segmenting the text part of the engineering archives of the housing construction project, the information processing unit of the text can be refined to better capture local semantic information. Compared with the entire document, the processing at the paragraph level can better understand the specific content and semantics of each part. Furthermore, each of the engineering archives of the housing construction project in the sequence of segment descriptions of the engineering archives of the housing construction project is respectively input into a segment content theme miner based on a large language model to obtain a sequence of segment content theme descriptions of the archives. That is, after the text part of the housing construction project engineering archive data is segmented, the segment content theme miner based on the large language model is used to perform content theme mining on each paragraph to deeply understand the semantics of the paragraph. The large language model is based on the context analysis method. By considering the keywords, phrases and sentence structures in the paragraph, it can accurately identify and extract the key topics and concepts in the paragraph, determine the core theme of each paragraph, and obtain the sequence of the content theme description of the archive data segment. In the housing construction project engineering archive data, the theme description provides fine-grained feature information of the housing construction project engineering, which helps the classifier to more accurately identify and classify documents and improve the accuracy of intelligent archiving.In particular, compared with traditional keyword matching methods, the large language model has learned a large amount of language patterns and knowledge through training, and can better understand complex semantic structures and accurately identify and extract the topic of a paragraph.

[0035] In particular, the S3 performs semantic encoding on each archival data segment content theme description in the sequence of the archival data segment content theme description to obtain a sequence of archival data segment content theme semantic encoding vectors. That is, in the technical solution of the present application, the archival data segment content theme description in the sequence of the archival data segment content theme description is semantically encoded to extract the contextual language semantic information in the archival data segment content theme description to obtain a sequence of archival data segment content theme semantic encoding vectors. The archival data segment content theme semantic encoding vector contains rich theme feature information and can better represent the theme and content of the paragraph. In a specific example, the Bert model can be used to semantically encode each archival data segment content theme description in the sequence of the archival data segment content theme description. The Bert model is a pre-trained model based on the Transformer architecture, which can capture the contextual information in the archival data segment content theme description.

[0036] In particular, the S4 performs semantic feature enhancement on the sequence of the archival material segment content theme semantic coding vectors to obtain a sequence of archival material segment content theme semantic enhancement coding vectors. It should be understood that, considering that the archival material segment content theme semantic coding vector contains the basic semantic information of the text, the main focus is on the semantic representation within a single paragraph, lacking the capture of the relationship between paragraphs. There may be inherent correlations between the archival material segment content themes. Therefore, in order to reveal the interaction and influence between the various content themes, thereby enhancing the expressive power of the features, the present application proposes a semantic enhancement method based on feature field effect distribution, which learns the contextual semantic associations between the various content themes by analyzing the semantic feature field effect distribution of the content themes of each archival material segment, thereby utilizing this contextual association relationship to strengthen the deep semantic understanding of the content themes of each archival material segment and enhance the semantic richness of the feature representation. In particular, in the technical solution of the present application, the semantic enhancement coding module based on sequence field effect distribution analysis is an advanced natural language processing architecture, which aims to perform refined feature extraction and semantic coding of the content theme of the archival material segment through multi-stage deep learning technology, so as to achieve in-depth analysis and semantic enhancement coding of the sequence of semantic coding vectors of the content theme of the archival material segment. Specifically, in the technical solution of the present application, if Figure 4As shown, the S4 includes: S41, based on the characteristic field effect distribution of the sequence of the archival data segment content subject semantic coding vectors, updating the sequence of the archival data segment content subject semantic coding vectors to obtain a sequence of updated archival data segment content subject semantic coding vectors; S42, performing bidirectional contextual semantic association enhancement on the sequence of updated archival data segment content subject semantic coding vectors to obtain a sequence of archival data segment content subject semantic enhancement coding vectors.

[0037] Specifically, the S41 updates the sequence of the archival data segment content theme semantic coding vectors based on the feature field effect distribution of the sequence of the archival data segment content theme semantic coding vectors to obtain a sequence of updated archival data segment content theme semantic coding vectors. In an embodiment of the present application, first, a semantic reconstructor is used to perform feature reconstruction on each archival data segment content theme semantic coding vector in the sequence of the archival data segment content theme semantic coding vectors to obtain a sequence of reconstructed archival data segment content theme semantic coding vectors; and a semantic reconstructor is used to perform domain transformation on each element in the sequence of the archival data segment content theme semantic coding vectors to generate a sequence of reconstructed archival data segment content theme semantic coding vectors. Through semantic reconstruction, the dimensionality of each archival data segment content theme semantic coding vector in the sequence of the archival data segment content theme semantic coding vectors can be unified, and at the same time, the representation ability of the vocabulary in the archival data segment content theme in a specific context is enhanced to be more in line with the meaning in the actual language environment. Next, the semantic depth implicit representation value of each reconstructed archival data segment content theme semantic coding vector in the sequence of the reconstructed archival data segment content theme semantic coding vector is calculated to obtain a sequence of reconstructed archival data segment content theme semantic depth implicit representation values; and based on the order of the sequence of the reconstructed archival data segment content theme semantic depth implicit representation values, the sequence of the archival data segment content theme semantic coding vector is reordered to obtain a sequence of updated archival data segment content theme semantic coding vectors. That is, in the technical solution of the present application, by calculating the semantic depth implicit representation value of these reconstructed feature vectors, the semantic depth implicit representation value reflects the information density or complexity carried by the vocabulary in the text, that is, by calculating the semantic depth implicit representation value of each reconstructed archival data segment content theme semantic coding vector to quantify the importance and influence of each vocabulary in the text paragraph where it is located, and according to the semantic depth implicit representation value corresponding to each archival data segment content theme semantic coding vector, the feature order in the sequence of the archival data segment content theme semantic coding vector is adjusted to ensure that key information is given priority in the subsequent contextual semantic association reinforcement coding process.

[0038] Among them, the specific steps of using a semantic reconstructor to reconstruct the features of each archival data segment content subject semantic coding vector in the sequence of the archival data segment content subject semantic coding vectors include: using a first weight parameter matrix and a second weight parameter matrix to perform continuous linear transformation on the archival data segment content subject semantic coding vector to obtain the reconstructed archival data segment content subject semantic coding vector.

[0039] More specifically, the specific steps of calculating the semantic depth implicit representation value of each reconstructed archival data segment content theme semantic coding vector in the sequence of the reconstructed archival data segment content theme semantic coding vector include: calculating the square of the Euclidean norm of the reconstructed archival data segment content theme semantic coding vector to obtain an initial semantic depth implicit representation value; calculating the ratio of the sum of the initial semantic depth implicit representation value and the difference between the initial semantic depth implicit representation value to obtain a semantic depth representation modulation coefficient; calculating the logarithmic function value of the semantic depth representation modulation coefficient with base 2 to obtain the reconstructed archival data segment content theme semantic depth implicit representation value.

[0040] Specifically, the S42 performs bidirectional contextual semantic association enhancement on the sequence of the updated archival data segment content theme semantic coding vectors to obtain the sequence of the archival data segment content theme semantic enhancement coding vectors. In an embodiment of the present application, first, the sequence of the updated archival data segment content theme semantic coding vectors is input into the first semantic encoder based on the forward LSTM model to obtain the sequence of the forward archival data segment content theme global semantic coding feature vectors; and the sequence of the updated archival data segment content theme semantic coding vectors is input into the first semantic encoder based on the backward LSTM model to obtain the sequence of the backward archival data segment content theme global semantic coding feature vectors. Here, a bidirectional long short-term memory (LSTM) model is used to encode the sequence of the updated archival data segment content theme semantic coding vectors from the forward and backward directions respectively. Among them, the forward LSTM captures the information flow pattern from the beginning of the sentence to the end of the sentence, while the backward LSTM operates in reverse to identify the dependency relationship from the end of the sentence to the beginning of the sentence. This dual-perspective encoding mechanism effectively captures the complex temporal structure in the archival data segment content theme, and uses it to enhance the semantic feature expression of each archival data segment content theme. Then, the sequence of the forward archival material segment content theme global semantic coding feature vectors and the sequence of the backward archival material segment content theme global semantic coding feature vectors are fused to obtain a sequence of archival material segment content theme semantic reinforcement coding vectors. By fusing the coding results of the bidirectional LSTM model to integrate the bidirectional contextual semantic information of the archival material segment content theme, a richer and more comprehensive sequence of archival material segment content theme semantic reinforcement coding vectors is formed, thereby achieving a deep understanding of the archival material segment content theme information.

[0041] Among them, the specific steps of fusing the sequence of the forward archival data segment content theme global semantic coding feature vectors and the sequence of the backward archival data segment content theme global semantic coding feature vectors include: using the first weighting parameter and the second weighting parameter to weighted modulate the sequence of the forward archival data segment content theme global semantic coding feature vectors and the sequence of the backward archival data segment content theme global semantic coding feature vectors respectively to obtain a sequence of weighted modulated forward archival data segment content theme global semantic coding feature vectors and a sequence of weighted modulated backward archival data segment content theme global semantic coding feature vectors; calculating the corresponding points between the sequence of weighted modulated forward archival data segment content theme global semantic coding feature vectors, the sequence of weighted modulated backward archival data segment content theme global semantic coding feature vectors and the bias item to obtain the sequence of archival data segment content theme semantic enhancement coding vectors.

[0042] In summary, in the above embodiment, the sequence of the archival material segment content subject semantic coding vectors is subjected to semantic feature enhancement to obtain the sequence of the archival material segment content subject semantic enhancement coding vectors, including: the sequence of the archival material segment content subject semantic coding vectors is subjected to semantic feature enhancement to obtain the sequence of the archival material segment content subject semantic enhancement coding vectors using the following semantic feature enhancement formula; wherein the semantic feature enhancement formula is:

[0043] T={w1,w2,...,w n}

[0044] h i =W1w i W2

[0045]

[0046] T order ={v1,v2,...,v n}

[0047]

[0048] T f =M m v forward +H m v backward +b m

[0049] Where T represents the sequence of the subject semantic coding vectors of the archive data segment content, w1, w2, w i and w nrepresents the first, second, i-th and n-th archival material segment content subject semantic coding vectors in the sequence of the archival material segment content subject semantic coding vectors, n is the number of the archival material segment content subject semantic coding vectors, W1 and W2 represent the first weight parameter matrix and the second weight parameter matrix, respectively, h i is the subject semantic coding vector of the i-th reconstructed archive data segment, ‖·‖ represents the Euclidean norm of the calculated vector, log(·) represents the logarithmic function with base 2, and e i represents the implicit representation value of the semantic depth of the content of the i-th reconstructed archive data segment, T order Represents a sequence of subject semantic encoding vectors of updated archive data segments, v1, v2, v i and v n respectively represent the first, second, i-th and n-th update archive data segment content subject semantic coding vectors in the sequence of the update archive data segment content subject semantic coding vectors, represents the forward LSTM model, represents the backward LSTM model, v forward represents the sequence of global semantic encoding feature vectors of the content theme of the forward archival data segment, v backward represents the sequence of global semantic encoding feature vectors of the content theme of the backward archival data segment, M m and H m denote the first weighting parameter and the second weighting parameter respectively, b m is the bias term, T f A sequence of encoding vectors for the subject semantics reinforcement of the content of the archival material segment.

[0050] In particular, the S5 cascades the sequence of the archival material segment content theme semantic reinforcement coding vectors to obtain the archival material global content theme semantic cascade coding vector. That is, through cascading, the local semantic information of each paragraph can be integrated into a global semantic representation, revealing the overall semantic framework and theme development context of the document, which helps to better understand the content and structure of the entire document, thereby improving the accuracy of the classifier, ensuring that the document is correctly classified, and improving the performance and efficiency of the building construction project engineering archive data management system.

[0051] In particular, the S6 determines the intelligent filing result based on the global content theme semantic cascade coding vector of the archival material, and based on the intelligent filing result, the engineering archival material of the housing construction project is filed. In a specific example of the present application, the global content theme semantic cascade coding vector of the archival material is input into the document intelligent filing module based on the classifier to obtain the intelligent filing result, and the intelligent filing result is the type label of the engineering archival material of the housing construction project. That is, in the scheme of the present application, the global content theme semantic cascade coding feature information of the archival material is used for classification processing, so as to automatically identify the engineering archival material of the housing construction project, so as to judge the type (such as construction documents, acceptance documents, technical documents, etc.) of the engineering archival material of the housing construction project. Then, and based on the intelligent filing result, the engineering archival material of the housing construction project is filed. Specifically, the specific steps of inputting the global content subject semantic cascade coding vector of the archival material into the document intelligent archiving module based on the classifier to obtain the intelligent archiving result include: using multiple fully connected layers of the classifier to fully connect the global content subject semantic cascade coding vector of the archival material to obtain a coding classification feature vector; and passing the coding classification feature vector through the Softmax classification function of the classifier to obtain the intelligent archiving result. In this way, not only the automation of the archival material archiving process is realized, but also the problems of inaccurate archiving classification and reliance on manual labor existing in the prior art are effectively solved, and the efficiency and accuracy of the management of engineering archival materials of housing construction projects are improved.

[0052] Preferably, inputting the global content subject semantic cascade coding vector of the archive material into a document intelligent archiving module based on a classifier to obtain the intelligent archiving result comprises:

[0053] Based on the association value and distance value of the eigenvalue of the archival material global content theme semantic cascade coding vector, the archival material global content theme semantic cascade coding association response matrix and the archival material global content theme semantic cascade coding distance response matrix are determined, namely:

[0054] M1(i,j)=v i ×vj

[0055]

[0056] where v i ,v j ∈V∈R 1×L , V is the archival material global content theme semantic cascade coding vector, L is the length of the archival material global content theme semantic cascade coding vector, M1 is the archival material global content theme semantic cascade coding association response matrix, M2 is the archival material global content theme semantic cascade coding distance response matrix;

[0057] Performing matrix multiplication of the archival material global content theme semantic cascade coding vector and the archival material global content theme semantic cascade coding association response matrix to obtain the archival material global content theme semantic cascade coding association response vector;

[0058] Performing matrix multiplication of the archival material global content theme semantic cascade coding distance response matrix and the transposed vector of the archival material global content theme semantic cascade coding vector to obtain the archival material global content theme semantic cascade coding distance response vector;

[0059] After performing a dot product of the archival material global content theme semantic cascade coding association response matrix and the archival material global content theme semantic cascade coding distance response matrix, perform a matrix multiplication with the transposed vector of the archival material global content theme semantic cascade coding vector to obtain the archival material global content theme semantic cascade coding logical bias vector;

[0060] The archival material global content theme semantic cascade coding association response vector, the archival material global content theme semantic cascade coding distance response vector and the archival material global content theme semantic cascade coding logical offset vector are subjected to point calculation to obtain an optimized archival material global content theme semantic cascade coding vector; and

[0061] The optimized archival material global content subject semantic cascade coding vector is input into the classifier-based document intelligent archiving module to obtain the intelligent archiving result.

[0062] Here, the optimized archival material global content topic semantic cascade coding vector V′ is expressed as:

[0063]

[0064] In V1∈R 1×L , and V2∈R L×1 In the case of:

[0065]

[0066] Among them, V T is the transposed vector of the archival material global content theme semantic cascade coding vector, V1 is the response vector obtained by adding the archival material global content theme semantic cascade coding associated response vector and the archival material global content theme semantic cascade coding distance response vector, V2 is the archival material global content theme semantic cascade coding logical bias vector, is added by position, is matrix multiplication, ⊙ is positional multiplication, and V′ is the optimized archival material global content theme semantic cascade coding vector.

[0067] That is, each archival data segment content theme semantic coding vector in the sequence of archival data segment content theme semantic coding vectors respectively represents the text semantic coding features of each archival data segment content theme description in the sequence of archival data segment content theme descriptions. Taking into account the semantic sparsity of the text theme distribution between each archival data segment content theme description in the sequence of archival data segment content theme descriptions, when the sequence of archival data segment content theme semantic coding vectors is input into the semantic feature enhancement based on sequence field effect distribution analysis, the sequence of archival data segment content theme semantic enhancement coding vectors as a whole has a large area of ​​simple repetition as a content theme cascade distribution, which affects the content theme semantic logical dependency and reduces the accuracy of the classification results based on the content theme semantic distribution.

[0068] Therefore, by taking the self-correlation matrix and self-distance matrix of the global content theme semantic cascade coding vector of the archival material as the statistical reference-free distribution response framework of the global content theme semantic cascade coding vector of the archival material, based on the retrieval-enhanced reverse response construction for the global content theme semantic cascade coding vector of the archival material, the simple repetition of feature distribution is avoided, and by ensuring the intrinsic mapping logic of the global content theme semantic cascade coding vector of the archival material based on the retrieval-response context, the surface combinatorial mapping of the global content theme semantic cascade coding vector of the archival material is avoided, so as to realize the logical dependency mapping of the global content theme semantic cascade coding vector of the archival material to the classification target domain while maintaining the intuitive response relevance, thereby improving the accuracy of the intelligent archiving result obtained by inputting the global content theme semantic cascade coding vector of the archival material into the document intelligent archiving module based on the classifier.

[0069] In summary, the method for managing the archival materials of the building construction project according to the embodiment of the present application is explained, which performs segment-granular topic mining on the engineering archival materials of the building construction project uploaded by the user by utilizing the natural language processing technology based on deep learning to extract the main content of the archival materials, and captures the semantic feature representation of each paragraph theme by semantically encoding the obtained each paragraph theme, and further, based on the contextual semantic association structure between each paragraph theme, performs semantic enhancement processing on it to achieve a deep semantic understanding of the global context information of the engineering archival materials of the building construction project, and based on this, intelligently judges and assigns specific type labels of the archival materials, thereby realizing the automated intelligent archiving of the engineering archival materials of the building construction project. In this way, the efficiency and accuracy of the management of the engineering archival materials of the building construction project can be effectively improved, human intervention can be reduced, and the risk of misjudgment and omission can be reduced.

[0070] Furthermore, a building construction project engineering archive data management system is also provided.

[0071] Figure 4 FIG. 1 is a block diagram of a building construction project engineering archive data management system according to an embodiment of the present application. Figure 4 As shown, according to the housing construction project engineering archive data management system 300 of the embodiment of the present application, it includes: a housing construction project engineering archive data acquisition module 310, which is used to acquire the housing construction project engineering archive data uploaded by the user; an archive data theme mining module 320, which is used to perform content theme mining based on segment granularity on the housing construction project engineering archive data to obtain a sequence of archive data segment content theme descriptions; an archive data segment content theme semantic encoding module 330, which is used to perform semantic encoding on each archive data segment content theme description in the sequence of archive data segment content theme descriptions to obtain an archive data segment content theme semantic encoding vector The archival material segment content subject semantic enhancement module 340 is used to enhance the semantic features of the sequence of archival material segment content subject semantic encoding vectors to obtain a sequence of archival material segment content subject semantic enhancement encoding vectors; the cascade module 350 is used to cascade the sequence of archival material segment content subject semantic enhancement encoding vectors to obtain archival material global content subject semantic cascade encoding vectors; the intelligent archiving module 360 ​​is used to determine the intelligent archiving result based on the archival material global content subject semantic cascade encoding vector, and archive the building construction project engineering archival material based on the intelligent archiving result.

[0072] As described above, the housing construction project engineering archive data management system 300 according to the embodiment of the present application can be implemented in various wireless terminals, such as a server with a housing construction project engineering archive data management algorithm. In a possible implementation, the housing construction project engineering archive data management system 300 according to the embodiment of the present application can be integrated into the wireless terminal as a software module and / or a hardware module. For example, the housing construction project engineering archive data management system 300 can be a software module in the operating system of the wireless terminal, or it can be an application developed for the wireless terminal; of course, the housing construction project engineering archive data management system 300 can also be one of the many hardware modules of the wireless terminal.

[0073] Alternatively, in another example, the housing construction project engineering archive data management system 300 and the wireless terminal may also be separate devices, and the housing construction project engineering archive data management system 300 may be connected to the wireless terminal via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0074] The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A method for managing engineering archives of a building construction project, characterized in that: include: Obtain the engineering archives of housing construction projects uploaded by users; Performing segment-based content theme mining on the engineering archives of the housing construction project to obtain a sequence of archival archive segment content theme descriptions; Semantically encoding each archival data segment content subject description in the sequence of archival data segment content subject descriptions to obtain a sequence of archival data segment content subject semantic encoding vectors; The sequence of archival data segment content subject semantic coding vectors is enhanced with semantic features to obtain a sequence of archival data segment content subject semantic reinforcement coding vectors, including: based on the feature field effect distribution of the sequence of archival data segment content subject semantic coding vectors, the sequence of archival data segment content subject semantic coding vectors is updated to obtain a sequence of updated archival data segment content subject semantic coding vectors; the sequence of updated archival data segment content subject semantic coding vectors is enhanced with bidirectional context semantic association to obtain a sequence of archival data segment content subject semantic reinforcement coding vectors; Cascading the sequence of the archival material segment content subject semantics reinforcement coding vectors to obtain the archival material global content subject semantics cascade coding vector; Based on the global content theme semantic cascade coding vector of the archival material, an intelligent archiving result is determined, and based on the intelligent archiving result, the engineering archival material of the housing construction project is archived.

2. The method for managing construction project archives according to claim 1, characterized in that: The content theme mining based on segment granularity is performed on the engineering archives of the housing construction project to obtain a sequence of archival material segment content theme descriptions, including: Performing text recognition on the engineering archives of the housing construction project to obtain a text portion of the engineering archives of the housing construction project; Segment the text part of the building construction project engineering archive data to obtain a sequence of segment descriptions of the building construction project engineering archive data; Each of the building construction project engineering archive data in the sequence of segment descriptions of the building construction project engineering archive data is input into a segment content theme miner based on a large language model to obtain a sequence of segment content theme descriptions of the archive data.

3. The method for managing construction project archives according to claim 2, characterized in that: Based on the characteristic field effect distribution of the sequence of the archival data segment content subject semantic coding vectors, the sequence of the archival data segment content subject semantic coding vectors is updated to obtain an updated sequence of the archival data segment content subject semantic coding vectors, including: Using a semantic reconstructor, feature-reconstructing each archival data segment content subject semantic coding vector in the sequence of archival data segment content subject semantic coding vectors to obtain a sequence of reconstructed archival data segment content subject semantic coding vectors; Calculating the semantic depth implicit representation value of each reconstructed archival data segment content subject semantic encoding vector in the sequence of reconstructed archival data segment content subject semantic encoding vectors to obtain a sequence of reconstructed archival data segment content subject semantic depth implicit representation values; Based on the ordering of the sequence of the reconstructed archival data segment content subject semantic depth implicit representation values, the sequence of the archival data segment content subject semantic coding vectors is reordered to obtain a sequence of updated archival data segment content subject semantic coding vectors.

4. The method for managing construction project archives according to claim 3 is characterized in that: Using a semantic reconstructor to perform feature reconstruction on each archival material segment content subject semantic coding vector in the sequence of archival material segment content subject semantic coding vectors to obtain a sequence of reconstructed archival material segment content subject semantic coding vectors, including: The first weight parameter matrix and the second weight parameter matrix are used to perform continuous linear transformation on the subject semantic coding vector of the archive data segment content to obtain the reconstructed archive data segment content subject semantic coding vector.

5. The method for managing construction project archives according to claim 4, characterized in that: Calculating the semantic depth implicit representation value of each reconstructed archival data segment content subject semantic encoding vector in the sequence of the reconstructed archival data segment content subject semantic encoding vectors to obtain a sequence of reconstructed archival data segment content subject semantic depth implicit representation values, including: Calculating the square of the Euclidean norm of the subject semantic coding vector of the reconstructed archive data segment content to obtain an initial semantic depth implicit representation value; Calculating a ratio of a sum of the initial semantic depth implicit representation value and a difference of the initial semantic depth implicit representation value to obtain a semantic depth representation modulation coefficient; The logarithmic function value with base 2 of the semantic depth representation modulation coefficient is calculated to obtain the implicit representation value of the content theme semantic depth of the reconstructed archive data segment.

6. The method for managing construction project archives according to claim 5, characterized in that: Performing bidirectional context semantic association enhancement on the sequence of the updated archive data segment content subject semantic coding vectors to obtain the sequence of the archive data segment content subject semantic enhancement coding vectors, including: Inputting the sequence of the updated archive data segment content theme semantic encoding vectors into a first semantic encoder based on a forward LSTM model to obtain a sequence of forward archive data segment content theme global semantic encoding feature vectors; Inputting the sequence of the updated archive data segment content theme semantic encoding vectors into a first semantic encoder based on a backward LSTM model to obtain a sequence of backward archive data segment content theme global semantic encoding feature vectors; The sequence of the forward archival material segment content theme global semantic coding feature vectors and the sequence of the backward archival material segment content theme global semantic coding feature vectors are fused to obtain a sequence of archival material segment content theme semantic reinforcement coding vectors.

7. The method for managing construction project archives according to claim 6, characterized in that: The sequence of the forward archival material segment content theme global semantic coding feature vectors and the sequence of the backward archival material segment content theme global semantic coding feature vectors are merged to obtain a sequence of archival material segment content theme semantic reinforcement coding vectors, including: Using the first weighting parameter and the second weighting parameter, weighted modulate the sequence of the forward archival material segment content theme global semantic coding feature vectors and the sequence of the backward archival material segment content theme global semantic coding feature vectors to obtain a sequence of weighted modulated forward archival material segment content theme global semantic coding feature vectors and a sequence of weighted modulated backward archival material segment content theme global semantic coding feature vectors; The sequence of weighted modulated forward archival material segment content thematic global semantic coding feature vectors, the sequence of weighted modulated backward archival material segment content thematic global semantic coding feature vectors and the corresponding points between the bias items are calculated to obtain the sequence of archival material segment content thematic semantic enhancement coding vectors.

8. The method for managing construction project archives according to claim 7, characterized in that: Based on the global content subject semantic cascade coding vector of the archival material, an intelligent archiving result is determined, and based on the intelligent archiving result, the engineering archival material of the housing construction project is archived, including: Inputting the global content subject semantic cascade coding vector of the archive data into a document intelligent archiving module based on a classifier to obtain the intelligent archiving result, wherein the intelligent archiving result is a type label of the engineering archive data of the housing construction project; Based on the intelligent archiving result, the engineering archive data of the building construction project are archived.

9. A building construction project engineering archive data management system, characterized in that: include: The housing construction project engineering archive data acquisition module is used to acquire the housing construction project engineering archive data uploaded by the user; An archive data subject mining module is used to perform content subject mining based on segment granularity on the engineering archive data of the housing construction project to obtain a sequence of archival data segment content subject descriptions; An archive data segment content theme semantic coding module, used for semantically coding each archive data segment content theme description in the sequence of archive data segment content theme descriptions to obtain a sequence of archive data segment content theme semantic coding vectors; The archival material segment content theme semantic enhancement module is used to enhance the semantic features of the sequence of archival material segment content theme semantic coding vectors to obtain a sequence of archival material segment content theme semantic enhancement coding vectors; A cascading module, used for cascading the sequence of the archival material segment content subject semantics reinforcement coding vectors to obtain the archival material global content subject semantics cascade coding vector; The intelligent archiving module is used to determine the intelligent archiving result based on the global content theme semantic cascade coding vector of the archival material, and archive the engineering archival material of the housing construction project based on the intelligent archiving result.

Images