Project similarity evaluation method based on semantic comprehension and multi-dimensional feature comparison

Abstract

The invention discloses a project similarity evaluation method based on semantic comprehension and multi-dimensional feature comparison, and the method comprises the steps: obtaining the project text data of a current declaration project, carrying out the semantic analysis and semantic mode understanding processing, extracting the text semantic features according to the semantic mode, and obtaining the project semantic features; historical declaration project data is obtained to carry out key text feature extraction processing, a project management model is constructed according to key text features and corresponding project declaration stages, and project semantic features are input into the model to carry out multi-dimensional semantic feature comparison processing. And evaluating the item similarity according to a multi-dimensional feature comparison result to obtain a repeated item approval evaluation result of the item, performing feature independent comparison processing on the item semantic feature of each sub-item of the current declaration item, and performing independent evaluation on the sub-item similarity to obtain a sub-item similarity evaluation result of the current declaration item. The method and the device have the effects of effectively evaluating the item similarity and improving the accuracy of item similarity evaluation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of project evaluation, and in particular to a project similarity evaluation method based on semantic understanding and multi-dimensional feature comparison. BACKGROUND

[0002] At present, with the continuous deepening of power reform and the continuous development of science and technology, various types of scientific and technological research projects and scientific and technological achievement project evaluation are increasing, and the problem of repeated project establishment is becoming more and more serious. In the project declaration and management process, higher requirements are put forward for the detection and management of repeated project establishment.

[0003] The existing detection of repeated project establishment is usually through the way of project similarity evaluation to carry out repeated analysis and detection, and the similarity of project text is compared to evaluate and determine the repeated project establishment. However, a large amount of unstructured data of scientific and technological projects is difficult to effectively identify, and a lot of manpower and material resources are consumed in the similarity discrimination of the project to be established. It is difficult to evaluate the comprehensive competitiveness of the scientific and technological project declaration, and an effective evaluation system cannot be formed. Therefore, there is further optimization space for the project similarity evaluation method of repeated project establishment in the above related technologies. SUMMARY

[0004] In view of the problem that the project similarity evaluation of repeated project establishment in the prior art lacks an effective evaluation system, the present application provides a project similarity evaluation method based on semantic understanding and multi-dimensional feature comparison, which can construct a standard project similarity evaluation system, effectively evaluate the project similarity, and improve the accuracy of project similarity evaluation.

[0005] In the first aspect, the above application aims to achieve the following technical scheme: A project similarity evaluation method based on semantic understanding and multi-dimensional feature comparison, the method comprising: obtaining project text data of a current declared project and performing semantic analysis and semantic modal understanding processing on the project text data, extracting text semantic features from the semantic analysis results according to the semantic modal, and obtaining project semantic features; obtaining historical project declaration data of a historical declared project, performing key text feature extraction processing on the historical project declaration data, and constructing a project management model according to the key text features and corresponding project declaration stages; inputting the project semantic features into the project management model for multi-dimensional semantic feature comparison processing, evaluating the project similarity according to the multi-dimensional feature comparison results, and obtaining a project repeated project establishment evaluation result; performing feature individual comparison processing on the project semantic features of each sub-project of the current declared project, independently evaluating the sub-project similarity, and obtaining a sub-project similarity evaluation result of the current declared project.

[0006] The application can be further configured in a preferred example to: the project text data of the current declared project is obtained, and semantic analysis and semantic modal understanding processing are performed on the project text data, text semantic features are extracted from the semantic analysis results according to the semantic modal, and project semantic features are obtained, specifically including: The project text data of the current declared project is obtained, and data preprocessing is performed on the project text data, and semantic mapping is performed on the preprocessed project text vocabulary, and a text word vector of the project text data is obtained; The project text data is processed by structured disassembly, the semantic relationship between text words and sentences is analyzed, and the text word vector is processed by semantic modal analysis; The text word vector is processed by text semantic feature extraction according to the semantic modal and semantic relationship of the project text, and project semantic features are obtained.

[0007] The application can be further configured in a preferred example to: the historical project declaration data of the historical declared project is obtained, the key text features of the historical project declaration data are extracted, and a project management model is constructed according to the key text features and the corresponding project declaration stage, specifically including: The historical project declaration data and the historical project declaration results of the historical declared project are obtained, the key text of the historical project declaration data is analyzed according to the historical project declaration results, and the key text features of the historical project declaration data are extracted; According to the project declaration stage of the historical declared project, the key text features of the corresponding declaration stage are adjusted, and the text modal and the text semantic relationship of the adjusted key text are analyzed to obtain stage feature data; The stage feature data of different declaration stages is processed by feature comparison, the data similarity of different declaration stages is evaluated according to the comparison result, and the similarity constraint condition of the historical declared project is generated; The stage feature data is trained under the similarity constraint condition, and a project management model for evaluating the similarity of the project is constructed according to the training result.

[0008] The application can be further configured in a preferred example to: the project semantic features are input into the project management model for multi-dimensional semantic feature comparison processing, the similarity of the project is evaluated according to the multi-dimensional feature comparison result, and the project repeated project evaluation result is obtained, specifically including: According to the project type and the project declaration stage, the current declared project is matched with the constraint condition in the project management model, and the similarity evaluation constraint condition of the current declared project is obtained; The semantic features of the current application project are compared with the features of similar application projects in the model using a multi-dimensional feature comparison process. The project similarity coefficient of the current application project is calculated based on the comparison results. The matching status between the project similarity coefficient and the similarity assessment constraints is analyzed. Based on the matching status of the constraints, the probability of duplicate project approval for the current application is analyzed, and the evaluation result of duplicate project approval is obtained.

[0009] In a preferred embodiment, this application can be further configured as follows: the process of calculating the project similarity coefficient in the step of performing multi-dimensional feature comparison processing between the project semantic features of the current application project and the features of similar application projects in the model, and calculating the project similarity coefficient of the current application project based on the comparison results, specifically includes: The text similarity coefficient between text word vectors is calculated using formula (1), which is shown below: (1) in, The cosine similarity between text word vectors is represented by the text cosine similarity. These represent the text word vectors in the project. Indicates the number of word vectors in the text; The semantic similarity coefficient related to the semantic relationship of the text is calculated using formula (2), which is shown below: (2) Specifically, This represents the semantic similarity coefficient used to characterize the linear correlation of semantic relationships in texts. , These represent the text word vectors respectively. The corresponding expected value.

[0010] In a preferred embodiment, this application can be further configured as follows: performing separate feature comparison processing on the semantic features of each sub-project of the current application project, independently evaluating the sub-project similarity, and obtaining the sub-project similarity evaluation result of the current application project, specifically includes: The current application project is split into sub-projects, and the semantic features of each sub-project are divided simultaneously to obtain the sub-project semantic features of each sub-project. The semantic features of sub-items are compared separately, and the similarity between sub-items is independently evaluated based on the comparison results to obtain the sub-item similarity evaluation results. Based on the sub-project similarity assessment results, the duplication status of the sub-projects in the current application project is analyzed to obtain the data on duplicate sub-project establishment.

[0011] Secondly, the above-mentioned inventive objective of this application is achieved through the following technical solutions: A project similarity assessment system based on semantic understanding and multidimensional feature comparison, wherein the system is applied to the aforementioned project similarity assessment method based on semantic understanding and multidimensional feature comparison, and the system includes: The semantic understanding module is used to acquire the project text data of the current application project and perform semantic analysis and semantic modality understanding processing on the project text data. Based on the semantic modality, the semantic features of the text are extracted from the semantic analysis results to obtain the project semantic features. The model building module is used to obtain historical project application data of historical projects, perform key text feature extraction processing on the historical project application data, and build a project management model based on the key text features and the corresponding project application stage. The feature comparison module is used to input the semantic features of the project into the project management model for multi-dimensional semantic feature comparison processing, evaluate the project similarity based on the multi-dimensional feature comparison results, and obtain the evaluation results of project duplicate establishment. The similarity assessment module is used to perform feature comparison processing on the semantic features of each sub-project of the current application project, to independently assess the similarity of the sub-projects, and to obtain the sub-project similarity assessment result of the current application project.

[0012] Thirdly, the above-mentioned objectives of this application are achieved through the following technical solutions: A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the above-described item similarity assessment method based on semantic understanding and multidimensional feature comparison.

[0013] Fourthly, the above-mentioned objectives of this application are achieved through the following technical solutions: A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described item similarity assessment method based on semantic understanding and multidimensional feature comparison.

[0014] In summary, this application includes at least one of the following beneficial technical effects: 1. This application extracts textual semantic features from massive project text data through semantic analysis and semantic modal understanding of project application texts, reducing the workload of searching and reviewing project application documents. Using historical application projects as training samples, key textual features from the project application process are extracted for data training to construct a project management model. This helps establish a standard system for project evaluation. The project management model performs multi-dimensional semantic feature comparison processing on the semantic features of current application projects, analyzing project similarity from multiple perspectives. This allows for accurate assessment of duplicate project applications. Furthermore, the application refines the project similarity assessment by examining the sub-projects of the current application projects. The sub-project similarity assessment results are used to filter duplicate content in the current application projects, reducing errors in project application texts and improving the quality of project applications. Through dual similarity assessments of the overall project and detailed sub-project similarities, the accuracy of project similarity assessment is improved.

[0015] 2. This application obtains text word vectors by semantically mapping project text data, which helps to capture the semantic and grammatical relationships of words and improve the ability to capture semantic similarity of words. Through the structured decomposition of project text data, the semantic relationships between text words and sentences are analyzed, and then the semantic modality of text word vectors is learned and analyzed based on semantic relationships, thereby improving the learning and analysis capabilities of semantic deep learning and further improving the accuracy of semantic relationship and semantic modality analysis.

[0016] 3. This application uses historical application data and corresponding historical project application results as data training samples, and uses the results as a guide to reverse analyze the key data in project applications, thereby improving the accuracy of extracting key text features in the project application process. It also makes phased adjustments to key text features according to the project application stage, thereby conducting phased analysis of text modality and text semantic relationships at each application stage, reducing the workload of frequent text data searches. By generating corresponding similarity constraints through the similarity of data at different application stages, it limits the similarity threshold in the model construction process, thereby improving the fit between the project management model and the actual situation of project applications, improving the authenticity of project similarity assessment, and forming a unified standard for project application management and project similarity assessment. Attached Figure Description

[0017] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0018] Figure 1This is a flowchart illustrating the implementation of the project similarity assessment method based on semantic understanding and multi-dimensional feature comparison in this embodiment.

[0019] Figure 2 This is a flowchart illustrating the implementation of step S10 of the project similarity assessment method in this embodiment.

[0020] Figure 3 This is a flowchart illustrating the implementation of step S20 of the project similarity assessment method in this embodiment.

[0021] Figure 4 This is a flowchart illustrating the implementation of step S30 of the project similarity assessment method in this embodiment.

[0022] Figure 5 This is a flowchart illustrating the implementation of step S40 of the project similarity assessment method in this embodiment.

[0023] Figure 6 This is a structural block diagram of the project similarity assessment system based on semantic understanding and multi-dimensional feature comparison in this embodiment.

[0024] Figure 7 This is a schematic diagram of the internal structure of a computer device used to implement a project similarity assessment method. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0027] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0028] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0029] In one embodiment, such as Figure 1 As shown, this application discloses a method for evaluating item similarity based on semantic understanding and multi-dimensional feature comparison, which specifically includes the following steps: S10: Obtain the project text data of the current application project and perform semantic analysis and semantic modality understanding processing on the project text data. Extract text semantic features based on the semantic modality analysis results to obtain project semantic features.

[0030] Specifically, such as Figure 2 As shown, step S10 includes: S101: Obtain the project text data of the current application project, perform data preprocessing, perform semantic mapping on the preprocessed project text words, and obtain the text word vectors of the project text data.

[0031] Specifically, project text data is obtained through text scanning and processing. The project text is then converted according to a preset data format to obtain project text data that conforms to the model input format. The preprocessed project text words are then processed by high- and low-dimensional vector mapping. In this embodiment, the BERT model in deep learning is used to encode the project text words in both high and low dimensions using the Transformer architecture to generate text word vectors for the project text data.

[0032] S102: Perform structured decomposition of the project text data, analyze the semantic relationships between text words and sentences, and perform semantic modality analysis on the text word vectors.

[0033] Specifically, the text structure in the project text data is extracted using the Large Language Model (LLM) and the Dynamic Pydantic Model to form the text structure system of the current application project. The project text data is then decomposed into a structured set of decomposed text words and sentences. The BERT model is then used for unsupervised training of the text word and sentence set to capture the polysemy and contextual dependencies of the text words and sentences, thereby analyzing the semantic relationships between the text words and sentences. Furthermore, semantic modal understanding analysis is performed on the text word vectors, including multiple modalities such as text, image, and sound. The multimodal information data is then transformed into a unified semantic representation to facilitate cross-modal semantic understanding.

[0034] S103: Combining the semantic modalities and semantic relationships of the project text, perform text semantic feature extraction processing on the text word vectors to obtain the project semantic features.

[0035] Specifically, combining the semantic modalities and semantic relationships of the project text, a deep learning model is used to perform text semantic feature analysis and extraction on the text word vectors. In this embodiment, a convolutional neural network is used to extract features such as part of speech, syntactic structure, word order, and sentence dependency relationships from the text word vectors to obtain the project semantic features.

[0036] S20: Obtain historical project application data for historical projects, extract key text features from the historical project application data, and construct a project management model based on the key text features and the corresponding project application stage.

[0037] Specifically, such as Figure 3 As shown, step S20 includes: S201: Obtain historical project application data and results for historical projects, analyze key texts of project applications based on feedback from historical project application results, and extract key text features from historical project application data.

[0038] Specifically, historical application data and corresponding application results of past projects are obtained. The application results are then fed back into the project application text for key text analysis and extraction. Specifically, the data of projects that failed to pass the application and the data of projects that passed the application are used as a control group for comparative analysis. Based on the comparison results, the key texts that led to the failure of the application are extracted, and key text features are extracted.

[0039] S202: Based on the project application stages of historical applications, the key text features of the corresponding application stages are adjusted in stages, and the text modality and semantic relationship of the adjusted key texts are analyzed to obtain stage feature data.

[0040] Specifically, according to the project application stage, the key text features of the historical application projects are divided into stages to obtain a set of key text features for each stage. The dependency relationship of the set of key text features for each stage is analyzed by an LSTM model to adjust the data loss and out-of-order problems that occurred during the text feature division. An attention mechanism is introduced to highlight the key text features in the set of key text features for each stage, so that the text modality and text semantic relationship of the marked key text can be analyzed independently to obtain stage feature data.

[0041] S203: Perform feature comparison processing on the stage feature data of different application stages, evaluate the data similarity of different application stages based on the comparison results, and generate similarity constraints for historical application projects.

[0042] Specifically, the feature data of the same application document at different application stages are compared, and the text features under the same position or title are compared and analyzed to assess the data similarity at each application stage. If the comparison results are consistent, it means that the data similarity is high and the text has not been modified. If the data similarity is low, it means that the text has been modified or there are errors. The analysis of how much data similarity is required to cause the project application to fail is used to plan the data similarity threshold and generate similarity constraints for historical application projects.

[0043] S204: Train the stage feature data under similarity constraints, and construct a project management model for project similarity assessment based on the training results.

[0044] Specifically, under similarity constraints, the stage feature data is trained using the TensorFlow and PYtorch deep learning frameworks, and the semantic similarity of the stage feature data is quantitatively evaluated to construct a project management model for evaluating project similarity.

[0045] S30: Input the project's semantic features into the project management model for multi-dimensional semantic feature comparison processing, evaluate the project similarity based on the multi-dimensional feature comparison results, and obtain the project duplication assessment results.

[0046] Specifically, such as Figure 4 As shown, step S30 includes: S301: Based on the project type and project application stage, match the current application project with the constraints in the project management model to obtain the similarity assessment constraints for the current application project.

[0047] Specifically, the project type and application stage of the current application project are obtained, and constraints of the same type and application stage are found in the project management model. Constraint matching is performed on the current application project, and similarity assessment constraints are obtained based on the matching results.

[0048] S302: Perform multi-dimensional feature comparison processing between the semantic features of the current application project and the features of similar application projects in the model, and calculate the project similarity coefficient of the current application project based on the comparison results.

[0049] Specifically, under the constraint of similarity assessment, i.e., the similarity difference is within the constraint threshold range, the semantic features of the current application project are compared with the features of the application projects in the model using multiple dimensions. Based on the comparison results, the project similarity coefficient of the current application project is calculated. The expression for the project similarity coefficient in this embodiment is as follows: The text similarity coefficient between text word vectors is calculated using formula (1), which is shown below: (1) in, The cosine similarity between text word vectors is represented by the text cosine similarity. These represent the text word vectors in the project. This indicates the number of word vectors in the text.

[0050] The semantic similarity coefficient related to the semantic relationship of the text is calculated using formula (2), which is shown below: (2) Specifically, This represents the semantic similarity coefficient used to characterize the linear correlation of semantic relationships in texts. , These represent the text word vectors respectively. The corresponding expected value.

[0051] S303: Analyze the matching status between the project similarity coefficient and the similarity assessment constraints, analyze the probability of duplicate project approval based on the matching status of the constraints, and obtain the evaluation results of duplicate project approval.

[0052] Specifically, the matching status of the project similarity coefficient and the similarity assessment constraints is analyzed. If the matching status of the similarity assessment constraints is within the similarity threshold range, it is considered a successful match. If it exceeds the similarity threshold, it is considered a duplicate project. Based on the matching status of the constraints, the probability of duplicate project approval for the current application is analyzed. In this embodiment, the relationship between the project similarity coefficient and the preset similarity threshold is used to analyze duplicate project approval. If the preset similarity threshold is exceeded, it is considered a duplicate project approval, and the duplicate project approval assessment result is obtained.

[0053] S40: Perform feature comparison processing on the semantic features of each sub-project of the current application project, evaluate the sub-project similarity independently, and obtain the sub-project similarity evaluation result of the current application project.

[0054] Specifically, such as Figure 5 As shown, step S40 includes: S401: Perform sub-project splitting on the current application project, and simultaneously divide the project semantic features of each sub-project to obtain the sub-project semantic features of each sub-project.

[0055] Specifically, the current application project is split into sub-projects according to the sub-project title. Based on the sub-project splitting results, the semantic feature data of the sub-projects are simultaneously divided to obtain the semantic features of each sub-project.

[0056] S402: Perform feature comparison processing on the semantic features of sub-items between sub-items, and independently evaluate the similarity between sub-items based on the comparison results to obtain the sub-item similarity evaluation results.

[0057] Specifically, the semantic features of sub-items are compared separately, such as calculating the text similarity coefficient and semantic similarity coefficient between the semantic features of sub-items, and analyzing the degree of similarity between words and sentences of sub-items. The higher the similarity coefficient, the greater the similarity. Thus, the similarity between sub-items is evaluated independently, and the sub-item similarity evaluation results are obtained.

[0058] S403: Based on the sub-project similarity assessment results, analyze the duplication status of sub-projects in the current application projects to obtain data on duplicate sub-project approvals.

[0059] Specifically, based on the sub-project similarity assessment results, the duplication status of sub-projects in the current application projects is analyzed. For example, by calculating the project similarity coefficient between two sub-projects, the higher the similarity coefficient, the greater the probability of duplication. When the similarity coefficient reaches a preset threshold, the sub-projects are identified as duplicate projects, thus obtaining the data on duplicate sub-projects.

[0060] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0061] In one embodiment, a project similarity assessment system based on semantic understanding and multi-dimensional feature comparison is provided. This system corresponds one-to-one with the project similarity assessment method based on semantic understanding and multi-dimensional feature comparison described in the above embodiments. For example... Figure 6 As shown, this project similarity assessment system based on semantic understanding and multi-dimensional feature comparison includes a semantic understanding module, a model building module, a feature comparison module, and a similarity assessment module. Detailed descriptions of each functional module are as follows: The semantic understanding module is used to acquire the project text data of the current application project and perform semantic analysis and semantic modality understanding processing on the project text data. Based on the semantic modality, the semantic features of the text are extracted from the semantic analysis results to obtain the project semantic features.

[0062] The model building module is used to obtain historical project application data, extract key text features from the historical project application data, and build a project management model based on the key text features and the corresponding project application stage.

[0063] The feature comparison module is used to input the semantic features of the project into the project management model for multi-dimensional semantic feature comparison processing, evaluate the similarity of projects based on the multi-dimensional feature comparison results, and obtain the evaluation results of duplicate project initiation.

[0064] The similarity assessment module is used to perform feature comparison processing on the semantic features of each sub-project of the current application project, to independently assess the similarity of the sub-projects, and to obtain the sub-project similarity assessment result of the current application project.

[0065] Preferably, the semantic understanding module specifically includes: The vector mapping submodule is used to obtain the project text data of the current application project, perform data preprocessing, and perform semantic mapping on the preprocessed project text words to obtain the text word vectors of the project text data.

[0066] The modality analysis submodule is used to perform structured decomposition of project text data, analyze the semantic relationships between text words and sentences, and perform semantic modality analysis on text word vectors.

[0067] The feature extraction submodule is used to combine the semantic modalities and semantic relationships of the project text to perform text semantic feature extraction on the text word vectors, thereby obtaining the project semantic features.

[0068] Preferably, the model building module specifically includes: The key feature extraction submodule is used to obtain historical project application data and results, analyze key texts of project applications based on historical application results, and perform key text feature extraction processing on historical project application data.

[0069] The stage feature analysis submodule is used to adjust the key text features of the corresponding application stage according to the application stage of historical application projects, and analyze the text modality and semantic relationship of the adjusted key text to obtain stage feature data.

[0070] The constraint condition submodule is used to perform feature comparison processing on the stage feature data of different application stages, evaluate the data similarity of different application stages based on the comparison results, and generate similarity constraint conditions for historical application projects.

[0071] The model building submodule is used to train stage feature data under similarity constraints, and to build a project management model for project similarity assessment based on the training results.

[0072] Preferably, the feature comparison module specifically includes: The condition matching submodule is used to match the current application project with the constraints in the project management model based on the project type and the stage of the application, so as to obtain the similarity assessment constraints of the current application project.

[0073] The coefficient calculation submodule is used to perform multi-dimensional feature comparison processing between the semantic features of the current application project and the features of similar application projects in the model, and calculate the project similarity coefficient of the current application project based on the comparison results.

[0074] The project evaluation submodule is used to analyze the matching status between the project similarity coefficient and the similarity evaluation constraints. Based on the matching status of the constraints, it analyzes the probability of duplicate project approval for the current application and obtains the project duplicate approval evaluation results.

[0075] Preferably, the item similarity coefficient calculation process in the coefficient calculation submodule specifically includes: The text similarity coefficient between text word vectors is calculated using formula (1), which is shown below: (1) in, The cosine similarity between text word vectors is represented by the text cosine similarity. These represent the text word vectors in the project. This indicates the number of word vectors in the text.

[0076] The semantic similarity coefficient related to the semantic relationship of the text is calculated using formula (2), which is shown below: (2) Specifically, This represents the semantic similarity coefficient used to characterize the linear correlation of semantic relationships in texts. , These represent the text word vectors respectively. The corresponding expected value.

[0077] Preferably, the similarity assessment module specifically includes: The sub-project splitting module is used to split the current application project into sub-projects, and simultaneously divide the semantic features of each sub-project to obtain the sub-project semantic features of each sub-project.

[0078] The sub-item similarity assessment submodule is used to perform feature comparison processing on the semantic features of sub-items, and to independently evaluate the similarity between sub-items based on the comparison results, thereby obtaining the sub-item similarity assessment result.

[0079] The sub-project establishment evaluation sub-module is used to analyze the duplication status of sub-projects in the current application project based on the sub-project similarity evaluation results, and obtain the data on duplicate sub-project establishment.

[0080] Specific limitations regarding the item similarity assessment system based on semantic understanding and multidimensional feature comparison can be found in the limitations of the item similarity assessment method based on semantic understanding and multidimensional feature comparison mentioned above, and will not be repeated here. Each module in the aforementioned item similarity assessment system based on semantic understanding and multidimensional feature comparison can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0081] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 7 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and the database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database stores data related to item similarity assessment. The network interface communicates with external terminals via a network connection. When executed by the processor, the computer program implements an item similarity assessment method based on semantic understanding and multi-dimensional feature comparison.

[0082] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program being executed by a processor to implement the steps of a project similarity assessment method based on semantic understanding and multidimensional feature comparison.

[0083] Those skilled in the art will recognize that the units of the various examples described in connection with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of the invention.

[0084] In the embodiments provided by the present invention, it should be understood that the division of units is only a logical functional division. In actual implementation, there may be other division methods, such as multiple units can be combined into one unit, one unit can be split into multiple units, or some features can be ignored.

[0085] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0086] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0087] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A method for evaluating item similarity based on semantic understanding and multidimensional feature comparison, characterized in that, The method includes: Obtain the project text data of the current application project and perform semantic analysis and semantic modality understanding processing on the project text data. Extract text semantic features from the semantic analysis results based on the semantic modality to obtain project semantic features; Obtain historical project application data for historical projects, extract key text features from the historical project application data, and construct a project management model based on the key text features and the corresponding project application stage. The semantic features of the project are input into the project management model for multi-dimensional semantic feature comparison processing. The similarity of the projects is evaluated based on the multi-dimensional feature comparison results to obtain the evaluation results of duplicate project initiation. The semantic features of each sub-project of the current application project are compared separately, and the sub-project similarity is evaluated independently to obtain the sub-project similarity evaluation result of the current application project.

2. The project similarity assessment method based on semantic understanding and multi-dimensional feature comparison according to claim 1, characterized in that, The process of acquiring the project text data of the currently submitted project and performing semantic analysis and semantic modality understanding on the project text data, and extracting text semantic features based on the semantic modality analysis results to obtain project semantic features, specifically includes: The project text data of the current application project is obtained and preprocessed. Semantic mapping is performed on the preprocessed project text words to obtain the text word vectors of the project text data. The project text data is decomposed into a structured form, the semantic relationships between words and sentences are analyzed, and the word vectors of the text are subjected to semantic modality analysis. By combining the semantic modalities and semantic relationships of the project text, text semantic features are extracted from the word vectors of the text to obtain the project semantic features.

3. The project similarity assessment method based on semantic understanding and multi-dimensional feature comparison according to claim 1, characterized in that, The process of obtaining historical project application data, extracting key text features from the historical project application data, and constructing a project management model based on the key text features and the corresponding project application stage includes: Obtain historical project application data and results for historical projects, analyze key texts of project applications based on the historical project application results, and extract key text features from historical project application data. Based on the project application stages of historical applications, the key text features of the corresponding application stages are adjusted in stages, and the text modality and semantic relationship of the adjusted key texts are analyzed to obtain stage feature data. The feature comparison data of different application stages are processed, and the data similarity of different application stages is evaluated based on the comparison results to generate similarity constraints for the historical application projects. The stage feature data is trained under the aforementioned similarity constraints, and a project management model for project similarity assessment is constructed based on the training results.

4. The project similarity assessment method based on semantic understanding and multi-dimensional feature comparison according to claim 1, characterized in that, The process of inputting the project's semantic features into the project management model for multi-dimensional semantic feature comparison processing, evaluating project similarity based on the multi-dimensional feature comparison results, and obtaining the project duplication assessment result specifically includes: Based on the project type and project application stage, the current application project is matched with the constraints in the project management model to obtain the similarity evaluation constraints for the current application project. The semantic features of the current application project are compared with the features of similar application projects in the model using a multi-dimensional feature comparison process. The project similarity coefficient of the current application project is calculated based on the comparison results. The matching status between the project similarity coefficient and the similarity assessment constraints is analyzed. Based on the matching status of the constraints, the probability of duplicate project approval for the current application is analyzed, and the evaluation result of duplicate project approval is obtained.

5. The project similarity assessment method based on semantic understanding and multi-dimensional feature comparison according to claim 4, characterized in that, The process of comparing the semantic features of the current application project with the features of similar application projects in the model using multi-dimensional features, and calculating the project similarity coefficient of the current application project based on the comparison results, specifically includes: The text similarity coefficient between text word vectors is calculated using formula (1), which is shown below: （1） in, The cosine similarity between text word vectors is represented by the text cosine similarity. These represent the text word vectors in the project. Indicates the number of word vectors in the text; The semantic similarity coefficient related to the semantic relationship of the text is calculated using formula (2), which is shown below: （2） Specifically, This represents the semantic similarity coefficient used to characterize the linear correlation of semantic relationships in texts. , These represent the text word vectors respectively. The corresponding expected value.

6. The project similarity assessment method based on semantic understanding and multi-dimensional feature comparison according to claim 1, characterized in that, The process of performing individual feature comparison on the semantic features of each sub-project of the current application project, and independently evaluating the sub-project similarity to obtain the sub-project similarity evaluation result of the current application project, specifically includes: The current application project is split into sub-projects, and the semantic features of each sub-project are divided simultaneously to obtain the sub-project semantic features of each sub-project. The semantic features of sub-items are compared separately, and the similarity between sub-items is independently evaluated based on the comparison results to obtain the sub-item similarity evaluation results. Based on the sub-project similarity assessment results, the duplication status of the sub-projects in the current application project is analyzed to obtain the data on duplicate sub-project establishment.

7. A project similarity assessment system based on semantic understanding and multi-dimensional feature comparison, characterized in that, The system is applied to the item similarity assessment method based on semantic understanding and multidimensional feature comparison as described in any one of claims 1-6, and the system comprises: The semantic understanding module is used to acquire the project text data of the current application project and perform semantic analysis and semantic modality understanding processing on the project text data. Based on the semantic modality, the semantic features of the text are extracted from the semantic analysis results to obtain the project semantic features. The model building module is used to obtain historical project application data of historical projects, perform key text feature extraction processing on the historical project application data, and build a project management model based on the key text features and the corresponding project application stage. The feature comparison module is used to input the semantic features of the project into the project management model for multi-dimensional semantic feature comparison processing, evaluate the project similarity based on the multi-dimensional feature comparison results, and obtain the evaluation results of project duplicate establishment. The similarity assessment module is used to perform feature comparison processing on the semantic features of each sub-project of the current application project, to independently assess the similarity of the sub-projects, and to obtain the sub-project similarity assessment result of the current application project.

8. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the project similarity assessment method based on semantic understanding and multidimensional feature comparison as described in any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the project similarity assessment method based on semantic understanding and multidimensional feature comparison as described in any one of claims 1 to 6.

Images