Smart library reading promotion management method based on knowledge graph

Abstract

The invention relates to the technical field of promotion management, and provides an intelligent library reading promotion management method based on a knowledge graph, and the method comprises the following steps: S1, obtaining library collection resources, user behaviors and scene data, constructing the knowledge graph, and providing a semantic association basis and intelligent matching between book resources and user demands for intelligent recommendation; s2, dynamically adjusting the knowledge association weight, and improving the popularization precision, the adaptive ability and the intelligent decision of the popularization strategy for the system; and S3, analyzing the multi-dimensional behavior data of the user, mining hidden interests of the user, depicting an interest portrait for the user, and dynamically updating in real time. Through the parallelism and randomness of the quantum technology, the response ability of the strategy to sudden hotspots and user demand changes is improved, the reading promotion strategy is continuously kept efficient in a dynamic environment, and the quality and effect of reading promotion are continuously improved by continuously optimizing the promotion strategy, so that a smart library is promoted to develop in a more intelligent and more adaptive direction.

Description

Technical Field

[0001] The present invention relates to the technical field of promotion management, and in particular to a knowledge graph-based smart library reading promotion management method. Background Art

[0002] In the field of smart libraries, existing reading promotion management methods are typically based on traditional machine learning or rule engines. These methods analyze users' explicit behavioral data to construct interest models and then combine them with the underlying connections in knowledge graphs to recommend resources. While these methods can achieve basic reading promotion functions, they have significant limitations in addressing users' implicit interests, dynamic interest evolution, and adapting to complex scenarios. First, they struggle to effectively capture user behavioral connections across platforms and time and space, and are unable to deeply explore implicit needs such as interdisciplinary interests. Consequently, recommendations are limited to areas corresponding to users' explicit behaviors and lack the ability to explore underlying interests. Second, during resource allocation optimization, traditional algorithms are inefficient when processing the high-dimensional connections in knowledge graphs and lack dynamic feedback mechanisms, making them difficult to adapt to real-time changes in user needs and promotion scenarios. Third, evaluation systems and policy update mechanisms are relatively static, unable to synchronize updates between knowledge graphs and algorithm parameters in real time. Furthermore, they lack the ability to integrate multimodal feedback data, resulting in poor adaptability and stability of promotion strategies in complex environments. Summary of the Invention

[0003] In response to the problems existing in the prior art, the purpose of the present invention is to provide a smart library reading promotion management method based on knowledge graph to solve the problems raised by the above background technology.

[0004] To achieve the above objectives, the present invention provides a method for managing reading promotion in a smart library based on a knowledge graph, comprising the following steps:

[0005] S1. Acquire library resources, user behavior, and scenario data to build a knowledge graph, providing a semantic association basis for intelligent recommendation and intelligent matching of book resources with user needs;

[0006] S2. Dynamically adjust the knowledge association weights to improve the system's promotion accuracy, adaptability, and intelligent decision-making for promotion strategies;

[0007] S3: Analyze users' multi-dimensional behavioral data, explore their implicit interests, create interest profiles for them, and update them dynamically in real time;

[0008] S4. Encode promotion strategies, generate optimal promotion combinations, and intelligently optimize resource allocation;

[0009] S5. Establish a feedback mechanism, dynamically adjust strategies based on feedback information, and conduct intelligent evaluation and iterative optimization of promotion effects;

[0010] S6. Build an evaluation system to update knowledge graphs and algorithm parameters in real time, enhancing intelligent self-learning capabilities for promotion strategies;

[0011] S7. Generate intelligent interactive solutions for AR scenes and dynamic knowledge maps based on knowledge graphs, providing users with intelligent and innovative promotion methods.

[0012] Preferably, in step S2, dynamically adjusting the knowledge association weight includes the following steps:

[0013] S21. Calculate the book knowledge entropy and user demand entropy, construct a multi-objective fitness function, and decompose the global entropy change of quantitative knowledge association. The formula is:

[0014] ;

[0015] Where, is the global entropy change settlement result of knowledge association, is a three-dimensional tensor The tensor information entropy of To measure the true correlation three-dimensional tensor the error associated with the model's predictions, is the influence weight, is the penalty factor, is the sparsity value that affects the weight;

[0016] S22. Utilize the quantum tunneling effect to correlate interdisciplinary correlation strengths and combine it with chaotic sequences to generate highly diverse initial weight solutions, break through local optimality, deeply explore the relationship between reading resources and users, and support precise and diverse reading promotion strategies. The formula is:

[0017] ;

[0018] Where, For users For resources The initial correlation degree, For users For resources The basic correlation strength, To measure the quantum tunneling temperature threshold that breaks through the local optimum, is a chaotic sequence;

[0019] S23. Construct a non-uniform temperature field driven mutation probability. Based on the weight configuration and the chaos factor of the solution, dynamically adjust the mutation probability to optimize the resource recommendation and reading promotion effect. The formula is:

[0020] ;

[0021] Where, The probability of mutation during the optimization of the reading promotion strategy, is the gradient difference between the current recommendation weight and the ideal optimal recommendation weight, is the position-dependent temperature field, is the coefficient of variation sensitivity;

[0022] S24. With the help of fiber bundle theory, through the mapping of Lie algebra and solution manifold and local optimization operator, the calculation process of reading promotion strategy is collaboratively optimized. The formula is:

[0023] ;

[0024] Where, Genetic Algorithm exist The weight matrix at time t, is the mapping relationship from Lie algebra to solution manifold, is a local optimization operator, Particle Swarm Optimization exist The weight matrix at the moment;

[0025] S25. Combining drift, diffusion, and jump terms that follow a Poisson distribution, we can characterize the continuous changes in knowledge entity associations over time. We also incorporate the impact of sudden hot events to accurately capture dynamic associations in reading promotion and adapt strategies. The formula is:

[0026] ;

[0027] Where, For users For resources The small change in the correlation is the drift term, and , for Moment User For resources The strength of association, is the time interval, For small time intervals, The diffusion coefficient is used to measure the random fluctuation degree of the association change of knowledge entities. To drive the correlation change, a small increment of random disturbance factors, To measure the jump coefficient of sudden events, is the small increment of the average number of sudden hot events per unit time. For users For resources The associated state quantity.

[0028] Preferably, in step S3, the evaluation and optimization includes the following steps:

[0029] S31. Integrate users' spatial and temporal behaviors across platforms, use tensor products to capture spatiotemporal coupling effects, construct high-dimensional interest representations, and accurately characterize user interests.

[0030] S32. Combining quantum entanglement, integrating basic interest state and co-occurrence reading state, combining interest energy level difference and quantum tunneling mechanism, we can explore users' hidden, cross-domain nonlinear interests and promote personalized reading for users. The formula is:

[0031] ;

[0032] Where, To hide the entangled state of interest, is the total number of related states such as basic interest states or co-occurring reading states involved in the calculation, To describe the quantum state representation of user basic interests, is the tensor product, For the co-occurrence reading state, is the quantized deformation of the Boltzmann distribution;

[0033] S33. Reduce the dimension of high-dimensional user behavior data and map it to an interpretable interest manifold, and dynamically adjust the compactness of different users and interest clusters. The formula is:

[0034] ;

[0035] Where, is the high-dimensional behavior data tensor of the object to be processed for dimensionality reduction, and To transform high-dimensional behavior data tensor and interest matrix Perform product operations of specific dimensions, To control the regularization coefficient of the degree of influence on the optimization objective, The curvature of the manifold to control the compactness of the cluster of interest Regularization term;

[0036] S34, integrating the spatiotemporal attention mechanism, dynamically adapts to the dynamic evolution of users' interest profiles based on their short-term hot spot responses and long-term reading preferences. The formula is:

[0037] ;

[0038] Where, is the fractional derivative that reflects the historical cumulative impact of the evolution of interest, is the user interest vector, is the interest attenuation coefficient, is the influence coefficient of the attention mechanism, and are query and investment matrices respectively, Associating tensors for users

[0039] S35. Verify the model error rate based on historical data, dynamically adjust the fuzzy rules and membership function parameters, and complete the intelligent update of the demand model.

[0040] Preferably, in step S31, accurately depicting the reader's interests includes the following steps:

[0041] S311. Collect spatial and temporal behaviors from multiple systems in the library, map all behaviors onto the behavior manifold, and describe the continuous behavior trajectory. The formula is:

[0042] ;

[0043] Where, For users To behavior The shortest behavior trajectory, To describe the curvature of the difficulty of behavior, is the intensity of the behavior, is a constraint condition;

[0044] S312. Introduce supersymmetric variables for spatial behavior, distinguish between real behavior and potential behavior, and use non-standard analysis to describe temporal behavior. Limit the graph to discrete time windows to capture the behavioral correlation of infinitesimal time swordsmen.

[0045] S313. Integrate all behavioral fragments to construct a high-dimensional interest representation to describe the user's interest direction. The formula is:

[0046] ;

[0047] Where, is the user's high-dimensional interest representation, is the total number of behavioral dimensions, For the The credibility weight of the behavior, For the A spatial behavior matrix, For the A timing behavior matrix, is the tensor product.

[0048] Preferably, in step S4, intelligent optimization is performed by encoding resource allocation with quantum neural network, combining the target tensor driven by knowledge graph, adapting the high-dimensional association of knowledge graph under the constraints of balanced error and distribution entropy, and completing efficient reading promotion. The formula is:

[0049] ;

[0050] Where, is the resource allocation matrix, Tensors and matrices related to the promotion strategy, Output a quantum-optimized resource allocation map for the quantum neural root network, is the target tensor, is the entropy weight coefficient, The entropy allocated for the resource.

[0051] Preferably, in step S5, the intelligent evaluation and iterative optimization includes the following steps:

[0052] S51. Encode user feedback into quantum state measurement results and use quantum state tomography to evaluate the promotion effect. The formula is:

[0053] ;

[0054] Where, To generalize the feedback density matrix of quantum state mixing, is the number of feedback types, is the feedback ground state, is the user behavior tensor Quantum measurement operators for projective measurements;

[0055] S52. Combined with feedback information, quantum stochastic micro equations are used to describe the dynamic adjustment of the strategy tensor. The formula is:

[0056] ;

[0057] Where, A policy tensor for adjusting promotion strategies over time Differential change, is the coefficient of natural attenuation of the strategy when there is no feedback, is the knowledge graph guide item, To simulate random fluctuations in promotion, In order to control the impact of random fluctuations on the strategy, is the feedback learning rate, is the feedback difference operator;

[0058] S53. Compare the matching degree between the current strategy and the knowledge graph, and use the quantum variational inequality to evaluate the optimality of the promotion effect. When the inequality holds, the current strategy is locally optimal, and the formula is:

[0059] ;

[0060] Where, An ideal strategy driven by knowledge graphs, is a set of strategies constrained by resources and channels. It is a composite operator that includes knowledge graph association, feedback error, and resource cost;

[0061] S54. Using knowledge graph association as a reward function, quantum reinforcement learning is used to complete the iterative optimization of the strategy. Through the superposition of quantum states, both popular promotion and hidden related promotion are explored simultaneously. The formula is:

[0062] ;

[0063] Where, and are the quantum strategy parameters before and after the update, is the learning rate, Knowledge Graph The quantum reward operator that drives the promotion strategy, is the strategy probability distribution.

[0064] Preferably, in step S6, updating the knowledge graph and algorithm parameters includes the following steps:

[0065] S61. Simultaneously measure the promotion effect and knowledge graph quality. Use quantum information entropy to construct a multi-dimensional evaluation index to evaluate the uncertainty of the promotion effect and measure the complexity of knowledge association. The formula is:

[0066] ;

[0067] Where, is the multidimensional evaluation index parameter, 、 and is the corresponding weight coefficient, The von Neumann entropy of the feedback density matrix, is the quantum entanglement entropy of the knowledge graph, is quantum Fisher information;

[0068] S62. Fusion of user feedback and promotion effects: With the help of quantum state evolution mechanism, the quantum master equation is used to fuse user feedback and promotion effects, and the knowledge graph is dynamically updated in real time, stably, and adaptively to feedback. The formula is:

[0069] ;

[0070] Where, For knowledge graph over time The evolution rate of The knowledge graph Lyapunov operator acts on the time The knowledge graph state at is the updated intensity coefficient, The quantum update unitary operator that adjusts the internal associations of the knowledge graph based on user behavior, To seek deviation operation;

[0071] S63. Dynamically optimize algorithm parameters based on evaluation indicators, use quantum meta-learning to quickly adapt to new tasks of smart library reading promotion, and use quantum regularization to prevent overfitting to ensure the adaptability of promotion effect and knowledge graph quality. The formula is:

[0072] ;

[0073] Where, is the optimal algorithm parameter obtained after optimization, To cover various tasks in the promotion of reading in smart libraries Based on distribution Take expectations, Update parameters for measuring meta-learning The loss function of the performance error of the algorithm in the reading promotion task is: is the regularization strength, A quantum regularization term is used to prevent overfitting in the optimization of reading promotion parameters.

[0074] S64. Process the knowledge graph from the perspectives of mining knowledge causal relationships, aligning multimodal information, and verifying strategy stability, strengthen the dynamic update capability of the knowledge graph, the multimodal fusion effect, and the strategy robustness, and continue to deepen the deep integration of quantum technology and library scenarios.

[0075] Preferably, in step S63, the meta-learning update parameters are used to quickly adapt to the new task of reading promotion, and the formula is:

[0076] ;

[0077] Where, To control the learning rate of the meta-learning update during the parameter update step size, is the loss function Meta-learning parameters gradient.

[0078] Preferably, in step S64, processing the knowledge graph includes the following steps:

[0079] S641. Distinguish the causal and false correlations between promotion strategies and user behaviors. Use quantum causal inference to identify the causal correlations between promotion strategies and user behaviors, and avoid false correlations interfering with updates. The formula is:

[0080] ;

[0081] Where, is the quantum causal coefficient, A quantum causal unitary operator that encodes causal propagation from strategy to behavior, To measure the quantum mutual information between strategy and behavior, For smart library reading promotion strategy, For user behavior;

[0082] S642. Through quantum state superposition, align user text, space, and time series multimodal feedback data, resolve modal conflicts, and accurately update the knowledge graph. The formula is:

[0083] ;

[0084] Where, To uniformly update the multimodal superposition state of the knowledge graph, The modal weights reflect the degree of influence of different modes when superimposed, To verify the multimodal feedback state and knowledge graph subspace Quantum exchange test of alignment;

[0085] S643. Use quantum random walks to simulate extreme scenarios and verify the robustness of the reading promotion strategy. This prevents the promotion strategy from being only adapted to conventional scenarios and ensures that the promotion strategy operates effectively in unpopular and sudden scenarios. The formula is:

[0086] ;

[0087] Where, is the robustness score for judging the stability of the strategy, Reading promotion strategies for smart libraries The adjoint operator of A quantum robust operator that imposes extreme scenario interference on the generalization strategy, is the worst-case density matrix, To perform trace operation.

[0088] The present invention provides a method for managing reading promotion in smart libraries based on knowledge graphs, which has the following beneficial effects:

[0089] 1. By integrating users' cross-platform spatial and temporal behaviors, using tensor products to capture the spatiotemporal coupling effect, and combining quantum entanglement, interest manifold dimensionality reduction, and spatiotemporal attention mechanisms, we can achieve in-depth mining and dynamic characterization of users' implicit interests. This can accurately identify users' unexpressed cross-domain interest preferences. The real-time updated interest portraits can adapt to the dynamic changes in readers' reading habits, enabling reading promotion to be upgraded from extensive recommendations based on explicit behaviors to precise matching based on implicit interests, significantly improving the relevance of recommendations and user satisfaction, and providing a more scientific decision-making basis for personalized reading promotion.

[0090] 2. By using quantum neural networks to encode resource allocation, combining knowledge graph target tensors to optimize resource configuration, and building a feedback iteration mechanism through quantum state tomography, stochastic differential equations, variational inequalities and reinforcement learning, intelligent optimization of resource allocation is achieved. It can quickly find the optimal promotion combination in complex knowledge graph associations, and dynamically adjust strategies based on user feedback. It can effectively balance the promotion of popular resources and the mining of unpopular knowledge, maximize the promotion effect under limited resources, and through the parallelism and randomness of quantum technology, improve the strategy's ability to respond to sudden hot spots and changes in user needs, so that the reading promotion strategy can continue to maintain high efficiency in a dynamic environment.

[0091] 3. By constructing a multidimensional evaluation system based on quantum information entropy, combined with quantum master equations, meta-learning and causal mining, cross-modal alignment, robustness verification and other technologies, real-time updates of knowledge graphs and algorithm parameters are achieved, and the intelligent self-learning ability of promotion strategies is improved. It can automatically identify the true causal relationship between promotion strategies and user behaviors, integrate multimodal feedback data, enhance the dynamics and accuracy of knowledge graphs, verify the stability of strategies by simulating extreme scenarios, ensure that promotion strategies operate effectively in various complex situations, enable smart library systems to autonomously adapt to the evolution of knowledge graphs and changes in user needs, continuously optimize promotion strategies, continuously improve the quality and effectiveness of reading promotion, and promote the development of smart libraries in a smarter and more adaptive direction. BRIEF DESCRIPTION OF THE DRAWINGS

[0092] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0093] Figure 1 A schematic diagram of the process framework of a knowledge graph-based smart library reading promotion management method provided in this application. DETAILED DESCRIPTION

[0094] The following embodiments of the present invention are described in further detail in conjunction with the accompanying drawings and examples. The following embodiments are only used to illustrate the present invention and are not intended to limit the scope of the present invention.

[0095] like Figure 1 As shown, this embodiment proposes a smart library reading promotion management method based on knowledge graph, including the following steps:

[0096] S1. Acquire library resources, user behavior, and scenario data to build a knowledge graph, providing a semantic association basis for intelligent recommendation and intelligent matching of book resources with user needs;

[0097] S2. Dynamically adjust the knowledge association weights to improve the system's promotion accuracy, adaptability, and intelligent decision-making for promotion strategies;

[0098] S3: Analyze users' multi-dimensional behavioral data, explore their implicit interests, create interest profiles for them, and update them dynamically in real time;

[0099] S4. Encode promotion strategies, generate optimal promotion combinations, and intelligently optimize resource allocation;

[0100] S5. Establish a feedback mechanism, dynamically adjust strategies based on feedback information, and conduct intelligent evaluation and iterative optimization of promotion effects;

[0101] S6. Build an evaluation system to update knowledge graphs and algorithm parameters in real time, enhancing intelligent self-learning capabilities for promotion strategies;

[0102] S7. Generate intelligent interactive solutions for AR scenes and dynamic knowledge maps based on knowledge graphs, providing users with intelligent and innovative promotion methods.

[0103] In this embodiment, in step S2, dynamically adjusting the knowledge association weight includes the following steps:

[0104] S21. Calculate the book knowledge entropy and user demand entropy, construct a multi-objective fitness function, and decompose the global entropy change of quantitative knowledge association. The formula is:

[0105] ;

[0106] Where, is the global entropy change settlement result of knowledge association, is a three-dimensional tensor The tensor information entropy of To measure the true correlation three-dimensional tensor the error associated with the model's predictions, is the influence weight, is the penalty factor, is the sparsity value that affects the weight;

[0107] S22. Utilize the quantum tunneling effect to correlate interdisciplinary correlation strengths and combine it with chaotic sequences to generate highly diverse initial weight solutions, break through local optimality, deeply explore the relationship between reading resources and users, and support precise and diverse reading promotion strategies. The formula is:

[0108] ;

[0109] Where, For users For resources The initial correlation degree, For users For resources The basic correlation strength, To measure the quantum tunneling temperature threshold that breaks through the local optimum, is a chaotic sequence;

[0110] S23. Construct a non-uniform temperature field driven mutation probability. Based on the weight configuration and the chaos factor of the solution, dynamically adjust the mutation probability to optimize the resource recommendation and reading promotion effect. The formula is:

[0111] ;

[0112] Where, The probability of mutation during the optimization of the reading promotion strategy, is the gradient difference between the current recommendation weight and the ideal optimal recommendation weight, is the position-dependent temperature field, is the coefficient of variation sensitivity;

[0113] S24. With the help of fiber bundle theory, through the mapping of Lie algebra and solution manifold and local optimization operator, the calculation process of reading promotion strategy is collaboratively optimized. The formula is:

[0114] ;

[0115] Where, Genetic Algorithm exist The weight matrix at time t, is the mapping relationship from Lie algebra to solution manifold, is a local optimization operator, Particle Swarm Optimization exist The weight matrix at the moment;

[0116] S25. Combining drift, diffusion, and jump terms that follow a Poisson distribution, we can characterize the continuous changes in knowledge entity associations over time. We also incorporate the impact of sudden hot events to accurately capture dynamic associations in reading promotion and adapt strategies. The formula is:

[0117] ;

[0118] Where, For users For resources The small change in the correlation is the drift term, and , for Moment User For resources The strength of association, is the time interval, For small time intervals, The diffusion coefficient is used to measure the random fluctuation degree of the association change of knowledge entities. To drive the correlation change, a small increment of random disturbance factors, To measure the jump coefficient of sudden events, is the small increment of the average number of sudden hot events per unit time. For users For resources The associated state quantity.

[0119] In this embodiment, in step S3, the evaluation and optimization includes the following steps:

[0120] S31. Integrate users' spatial and temporal behaviors across platforms, use tensor products to capture spatiotemporal coupling effects, construct high-dimensional interest representations, and accurately characterize user interests.

[0121] S32. Combining quantum entanglement, integrating basic interest state and co-occurrence reading state, combining interest energy level difference and quantum tunneling mechanism, we can explore users' hidden, cross-domain nonlinear interests and promote personalized reading for users. The formula is:

[0122] ;

[0123] Where, To hide the entangled state of interest, is the total number of related states such as basic interest states or co-occurring reading states involved in the calculation, To describe the quantum state representation of user basic interests, is the tensor product, For co-occurrence reading state, is the quantized deformation of the Boltzmann distribution;

[0124] S33. Reduce the dimension of high-dimensional user behavior data and map it to an interpretable interest manifold, and dynamically adjust the compactness of different users and interest clusters. The formula is:

[0125] ;

[0126] Where, is the high-dimensional behavior data tensor of the object to be processed for dimensionality reduction, and To transform high-dimensional behavior data tensors and interest matrices Perform product operations of specific dimensions, To control the regularization coefficient of the degree of influence on the optimization objective, The curvature of the manifold to control the compactness of the clusters of interest Regularization term;

[0127] S34, integrating the spatiotemporal attention mechanism, dynamically adapts to the dynamic evolution of the user's interest profile of short-term hot spot responses and long-term reading preferences. The formula is:

[0128] ;

[0129] Where, is the fractional derivative that reflects the historical cumulative impact of the evolution of interest, is the user interest vector, is the interest attenuation coefficient, is the influence coefficient of the attention mechanism, and are query and investment matrices respectively, Associating tensors for users

[0130] S35. Verify the model error rate based on historical data, dynamically adjust the fuzzy rules and membership function parameters, and complete the intelligent update of the demand model.

[0131] In this embodiment, in step S31, accurately depicting the reader's interests includes the following steps:

[0132] S311. Collect spatial and temporal behaviors from multiple systems in the library, map all behaviors onto the behavior manifold, and describe the continuous behavior trajectory. The formula is:

[0133] ;

[0134] Where, For users To behavior The shortest behavior trajectory, To describe the curvature of the difficulty of behavior, is the intensity of the behavior, is a constraint condition;

[0135] S312. Introduce supersymmetric variables for spatial behavior, distinguish between real behavior and potential behavior, and use non-standard analysis to describe temporal behavior. Limit the graph to discrete time windows to capture the behavioral correlation of infinitesimal time swordsmen.

[0136] S313. Integrate all behavioral fragments to construct a high-dimensional interest representation to describe the user's interest direction. The formula is:

[0137] ;

[0138] Where, is the user's high-dimensional interest representation, is the total number of behavioral dimensions, For the The credibility weight of the behavior, For the A spatial behavior matrix, For the A timing behavior matrix, is the tensor product.

[0139] Specifically, the present invention integrates users' cross-platform spatial and temporal behaviors, uses tensor products to capture the space-time coupling effect, and combines quantum entangled states, interest manifold dimensionality reduction and space-time attention mechanisms to achieve in-depth mining and dynamic characterization of users' implicit interests. It can accurately identify users' unexpressed cross-domain interest preferences, and the real-time updated interest portraits can adapt to the dynamic changes in readers' reading habits, so that reading promotion can be upgraded from extensive recommendations based on explicit behaviors to precise matching based on implicit interests, significantly improving the relevance of recommendations and user satisfaction, and providing a more scientific decision-making basis for personalized reading promotion.

[0140] In this embodiment, in step S4, intelligent optimization is performed by encoding resource allocation with quantum neural networks, combining the target tensor driven by the knowledge graph, adapting the high-dimensional association of the knowledge graph under the constraints of balanced error and distribution entropy, and completing efficient reading promotion. The formula is:

[0141] ;

[0142] Where, is the resource allocation matrix, Tensors and matrices related to the promotion strategy, Output a quantum-optimized resource allocation map for the quantum neural root network, is the target tensor, is the entropy weight coefficient, The entropy allocated for the resource.

[0143] In this embodiment, in step S5, the intelligent evaluation and iterative optimization includes the following steps:

[0144] S51. Encode user feedback into quantum state measurement results and use quantum state tomography to evaluate the promotion effect. The formula is:

[0145] ;

[0146] Where, To generalize the feedback density matrix of quantum state mixing, is the number of feedback types, is the feedback ground state, is the user behavior tensor Quantum measurement operators for projective measurements;

[0147] S52. Combined with feedback information, quantum stochastic micro equations are used to describe the dynamic adjustment of the strategy tensor. The formula is:

[0148] ;

[0149] Where, A policy tensor for adjusting promotion strategies over time Differential change, is the coefficient of natural attenuation of the strategy when there is no feedback, is the knowledge graph guide item, To simulate random fluctuations in promotion, In order to control the impact of random fluctuations on the strategy, is the feedback learning rate, is the feedback difference operator;

[0150] S53. Compare the matching degree between the current strategy and the knowledge graph, and use the quantum variational inequality to evaluate the optimality of the promotion effect. When the inequality holds, the current strategy is locally optimal, and the formula is:

[0151] ;

[0152] Where, An ideal strategy driven by knowledge graphs, is a set of strategies constrained by resources and channels. It is a composite operator that includes knowledge graph association, feedback error, and resource cost;

[0153] S54. Using knowledge graph association as a reward function, quantum reinforcement learning is used to complete the iterative optimization of the strategy. Through the superposition of quantum states, both popular promotion and hidden related promotion are explored simultaneously. The formula is:

[0154] ;

[0155] Where, and are the quantum strategy parameters before and after the update, is the learning rate, Knowledge Graph The quantum reward operator that drives the promotion strategy, is the strategy probability distribution.

[0156] Specifically, the present invention achieves intelligent optimization of resource allocation by encoding resource allocation with the help of quantum neural networks, combining the knowledge graph target tensor to optimize resource configuration, and constructing a feedback iteration mechanism through quantum state tomography, stochastic differential equations, variational inequalities and reinforcement learning. It can quickly find the optimal promotion combination in complex knowledge graph associations, and dynamically adjust the strategy according to user feedback. It can effectively balance the promotion of popular resources and the mining of unpopular knowledge, maximize the promotion effect under limited resources, and through the parallelism and randomness of quantum technology, improve the strategy's ability to respond to sudden hot spots and changes in user needs, so that the reading promotion strategy can continue to maintain high efficiency in a dynamic environment.

[0157] In this embodiment, in step S6, updating the knowledge graph and algorithm parameters includes the following steps:

[0158] S61. Simultaneously measure the promotion effect and knowledge graph quality. Use quantum information entropy to construct a multi-dimensional evaluation index to evaluate the uncertainty of the promotion effect and measure the complexity of knowledge association. The formula is:

[0159] ;

[0160] Where, is the multidimensional evaluation index parameter, 、 and is the corresponding weight coefficient, The von Neumann entropy of the feedback density matrix, is the quantum entanglement entropy of the knowledge graph, is quantum Fisher information;

[0161] S62. Fusion of user feedback and promotion effects: With the help of quantum state evolution mechanism, the quantum master equation is used to fuse user feedback and promotion effects, and the knowledge graph is dynamically updated in real time, stably, and adaptively to feedback. The formula is:

[0162] ;

[0163] Where, For knowledge graph over time The evolution rate of The knowledge graph Lyapunov operator acts on the time The knowledge graph state at is the updated intensity coefficient, The quantum update unitary operator that adjusts the internal associations of the knowledge graph based on user behavior, To seek deviation operation;

[0164] S63. Dynamically optimize algorithm parameters based on evaluation indicators, use quantum meta-learning to quickly adapt to new tasks of smart library reading promotion, and use quantum regularization to prevent overfitting to ensure the adaptability of promotion effect and knowledge graph quality. The formula is:

[0165] ;

[0166] Where, is the optimal algorithm parameter obtained after optimization, To cover various tasks in the promotion of reading in smart libraries Based on distribution Take expectations, Update parameters for measuring meta-learning The loss function of the performance error of the algorithm in the reading promotion task is: is the regularization strength, A quantum regularization term is used to prevent overfitting in the optimization of reading promotion parameters.

[0167] S64. Process the knowledge graph from the perspectives of mining knowledge causal relationships, aligning multimodal information, and verifying strategy stability, strengthen the dynamic update capability of the knowledge graph, the multimodal fusion effect, and the strategy robustness, and continue to deepen the deep integration of quantum technology and library scenarios.

[0168] In this embodiment, in step S63, meta-learning updates parameters to quickly adapt to the new task of reading promotion. The formula is:

[0169] ;

[0170] Where, To control the learning rate of the meta-learning update during the parameter update step size, is the loss function Meta-learning parameters gradient.

[0171] In this embodiment, in step S64, processing the knowledge graph includes the following steps:

[0172] S641. Distinguish the causal and false correlations between promotion strategies and user behaviors. Use quantum causal inference to identify the causal correlations between promotion strategies and user behaviors, and avoid false correlations interfering with updates. The formula is:

[0173] ;

[0174] Where, is the quantum causal coefficient, A quantum causal unitary operator that encodes causal propagation from strategy to behavior, To measure the quantum mutual information between strategy and behavior, To develop a reading promotion strategy for smart libraries, For user behavior;

[0175] S642. Through quantum state superposition, align user text, space, and time series multimodal feedback data, resolve modal conflicts, and accurately update the knowledge graph. The formula is:

[0176] ;

[0177] Where, To uniformly update the multimodal superposition state of the knowledge graph, The modal weights reflect the degree of influence of different modes when superimposed, To verify the multimodal feedback state and knowledge graph subspace Quantum exchange test of alignment;

[0178] S643. Use quantum random walks to simulate extreme scenarios and verify the robustness of the reading promotion strategy. This prevents the promotion strategy from being only adapted to conventional scenarios and ensures that the promotion strategy operates effectively in unpopular and sudden scenarios. The formula is:

[0179] ;

[0180] Where, is the robustness score for judging the stability of the strategy, Reading promotion strategies for smart libraries The adjoint operator of A quantum robust operator that imposes extreme scenario interference on the generalization strategy, is the worst-case density matrix, To perform trace operation.

[0181] Specifically, the present invention constructs a multidimensional evaluation system based on quantum information entropy, combines quantum master equations, meta-learning and causal mining, cross-modal alignment, robustness verification and other technologies, realizes real-time updating of knowledge graphs and algorithm parameters, improves the intelligent self-learning ability of promotion strategies, automatically identifies the true causal relationship between promotion strategies and user behaviors, integrates multimodal feedback data, enhances the dynamics and accuracy of knowledge graphs, verifies strategy stability by simulating extreme scenarios, ensures that promotion strategies operate effectively in various complex situations, enables smart library systems to autonomously adapt to the evolution of knowledge graphs and changes in user needs, continuously optimizes promotion strategies, continuously improves the quality and effectiveness of reading promotion, and promotes the development of smart libraries in a smarter and more adaptive direction.

[0182] The above embodiments are intended to illustrate the present invention only and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the embodiments, it should be understood by those skilled in the art that various combinations, modifications, or equivalent substitutions of the technical solutions of the present invention do not depart from the spirit and scope of the technical solutions of the present invention and should be encompassed by the scope of the claims of the present invention.

Claims

1. A method for reading promotion and management in a smart library based on knowledge graph, characterized in that: The following steps are involved: S1. Acquire library resources, user behavior, and scenario data to build a knowledge graph, providing a semantic association basis for intelligent recommendation and intelligent matching of book resources with user needs; S2. Dynamically adjust the knowledge association weights to improve the system's promotion accuracy, adaptability, and intelligent decision-making for promotion strategies; S3: Analyze users' multi-dimensional behavioral data, explore their implicit interests, create interest profiles for them, and update them dynamically in real time; S4. Encode promotion strategies, generate optimal promotion combinations, and intelligently optimize resource allocation; S5. Establish a feedback mechanism, dynamically adjust strategies based on feedback information, and conduct intelligent evaluation and iterative optimization of promotion effects; S6. Build an evaluation system to update knowledge graphs and algorithm parameters in real time, enhancing intelligent self-learning capabilities for promotion strategies; S7. Generate intelligent interactive solutions for AR scenes and dynamic knowledge maps based on knowledge graphs, providing users with intelligent and innovative promotion methods.

2. The method for promoting reading in a smart library based on knowledge graph according to claim 1 is characterized in that: In step S2, dynamically adjusting the knowledge association weight includes the following steps: S21. Calculate the book knowledge entropy and user demand entropy, construct a multi-objective fitness function, and decompose the global entropy change of quantitative knowledge association. The formula is: ； Where, is the global entropy change settlement result of knowledge association, is a three-dimensional tensor The tensor information entropy of To measure the true correlation three-dimensional tensor the error associated with the model's predictions, is the influence weight, is the penalty factor, is the sparsity value that affects the weight; S22. Utilize the quantum tunneling effect to correlate interdisciplinary correlation strengths and combine it with chaotic sequences to generate highly diverse initial weight solutions, break through local optimality, deeply explore the relationship between reading resources and users, and support precise and diverse reading promotion strategies. The formula is: ； Where, For users For resources The initial correlation degree, For users For resources The basic correlation strength, To measure the quantum tunneling temperature threshold that breaks through the local optimum, is a chaotic sequence; S23. Construct a non-uniform temperature field driven mutation probability. Based on the weight configuration and the chaos factor of the solution, dynamically adjust the mutation probability to optimize the resource recommendation and reading promotion effect. The formula is: ； Where, The probability of mutation during the optimization of the reading promotion strategy, is the gradient difference between the current recommendation weight and the ideal optimal recommendation weight, is the position-dependent temperature field, is the coefficient of variation sensitivity; S24. With the help of fiber bundle theory, through the mapping of Lie algebra and solution manifold and local optimization operator, the calculation process of reading promotion strategy is collaboratively optimized. The formula is: ； Where, Genetic Algorithm exist The weight matrix at time t, is the mapping relationship from Lie algebra to solution manifold, is a local optimization operator, Particle Swarm Optimization exist The weight matrix at the moment; S25. Combining drift, diffusion, and jump terms that follow a Poisson distribution, we can characterize the continuous changes in knowledge entity associations over time. We also incorporate the impact of sudden hot events to accurately capture dynamic associations in reading promotion and adapt strategies. The formula is: ； Where, For users For resources The small change in the correlation is the drift term, and , for Moment User For resources The strength of association, is the time interval, For small time intervals, The diffusion coefficient is used to measure the random fluctuation degree of the association change of knowledge entities. To drive the correlation change, a small increment of random disturbance factors, To measure the jump coefficient of sudden events, is the small increment of the average number of sudden hot events per unit time. For users For resources The associated state quantity.

3. The method for promoting reading in a smart library based on knowledge graph according to claim 1 is characterized in that: In step S3, the evaluation and optimization includes the following steps: S31. Integrate users' spatial and temporal behaviors across platforms, use tensor products to capture spatiotemporal coupling effects, construct high-dimensional interest representations, and accurately characterize user interests. S32. Combining quantum entanglement, integrating basic interest state and co-occurrence reading state, combining interest energy level difference and quantum tunneling mechanism, we can explore users' hidden, cross-domain nonlinear interests and promote personalized reading for users. The formula is: ； Where, To hide the entangled state of interest, is the total number of related states such as basic interest states or co-occurring reading states involved in the calculation, To describe the quantum state representation of user basic interests, is the tensor product, For co-occurrence reading state, is the quantized deformation of the Boltzmann distribution; S33. Reduce the dimension of high-dimensional user behavior data and map it to an interpretable interest manifold, and dynamically adjust the compactness of different users and interest clusters. The formula is: ； Where, is the high-dimensional behavior data tensor of the object to be processed for dimensionality reduction, and To transform high-dimensional behavior data tensors and interest matrices Perform product operations of specific dimensions, To control the regularization coefficient of the degree of influence on the optimization objective, The curvature of the manifold to control the compactness of the clusters of interest Regularization term; S34, integrating the spatiotemporal attention mechanism, dynamically adapts to the dynamic evolution of users' interest profiles based on their short-term hot spot responses and long-term reading preferences. The formula is: ； Where, is the fractional derivative that reflects the historical cumulative impact of the evolution of interest, is the user interest vector, is the interest attenuation coefficient, is the influence coefficient of the attention mechanism, and are query and investment matrices respectively, Associating tensors for users S35. Verify the model error rate based on historical data, dynamically adjust the fuzzy rules and membership function parameters, and complete the intelligent update of the demand model.

4. A method for promoting reading in a smart library based on knowledge graph according to claim 3, characterized in that: In step S31, accurately depicting the reader's interests includes the following steps: S311. Collect spatial and temporal behaviors from multiple systems in the library, map all behaviors onto the behavior manifold, and describe the continuous behavior trajectory. The formula is: ； Where, For users To behavior The shortest behavior trajectory, To describe the curvature of the difficulty of behavior, is the intensity of the behavior, is a constraint condition; S312. Introduce supersymmetric variables for spatial behavior, distinguish between real behavior and potential behavior, and use non-standard analysis to describe temporal behavior. Limit the graph to discrete time windows to capture the behavioral correlation of infinitesimal time swordsmen. S313. Integrate all behavioral fragments to construct a high-dimensional interest representation to describe the user's interest direction. The formula is: ； Where, is the user's high-dimensional interest representation, is the total number of behavioral dimensions, For the The credibility weight of the behavior, For the A spatial behavior matrix, For the A timing behavior matrix, is the tensor product.

5. The method for promoting reading in a smart library based on knowledge graph according to claim 1 is characterized in that: In step S4, intelligent optimization uses quantum neural network encoding resource allocation, combined with the target tensor driven by the knowledge graph, to adapt the high-dimensional association of the knowledge graph under the constraints of balanced error and distribution entropy to achieve efficient reading promotion. The formula is: ； Where, is the resource allocation matrix, Tensors and matrices related to the promotion strategy, Output a quantum-optimized resource allocation map for the quantum neural root network, is the target tensor, is the entropy weight coefficient, The entropy allocated for the resource.

6. The method for promoting reading in a smart library based on knowledge graph according to claim 4 is characterized in that: In step S5, the intelligent evaluation and iterative optimization includes the following steps: S51. Encode user feedback into quantum state measurement results and use quantum state tomography to evaluate the promotion effect. The formula is: ； Where, To generalize the feedback density matrix of quantum state mixing, is the number of feedback types, is the feedback ground state, is the user behavior tensor Quantum measurement operators for projective measurements; S52. Combined with feedback information, quantum stochastic micro equations are used to describe the dynamic adjustment of the strategy tensor. The formula is: ； Where, A policy tensor for adjusting promotion strategies over time Differential change, is the coefficient of natural attenuation of the strategy when there is no feedback, is the knowledge graph guide item, To simulate random fluctuations in promotion, In order to control the impact of random fluctuations on the strategy, is the feedback learning rate, is the feedback difference operator; S53. Compare the matching degree between the current strategy and the knowledge graph, and use the quantum variational inequality to evaluate the optimality of the promotion effect. When the inequality holds, the current strategy is locally optimal, and the formula is: ； Where, An ideal strategy driven by knowledge graphs, is a set of strategies constrained by resources and channels. It is a composite operator that includes knowledge graph association, feedback error, and resource cost; S54. Using knowledge graph association as a reward function, quantum reinforcement learning is used to complete the iterative optimization of the strategy. Through the superposition of quantum states, both popular promotion and hidden related promotion are explored simultaneously. The formula is: ； Where, and are the quantum strategy parameters before and after the update, is the learning rate, Knowledge Graph The quantum reward operator that drives the promotion strategy, is the strategy probability distribution.

7. The method for managing reading promotion in a smart library based on knowledge graph according to claim 6 is characterized in that: In step S6, updating the knowledge graph and algorithm parameters includes the following steps: S61. Simultaneously measure the promotion effect and knowledge graph quality. Use quantum information entropy to construct a multi-dimensional evaluation index to evaluate the uncertainty of the promotion effect and measure the complexity of knowledge association. The formula is: ； Where, is the multidimensional evaluation index parameter, 、 and is the corresponding weight coefficient, The von Neumann entropy of the feedback density matrix, is the quantum entanglement entropy of the knowledge graph, is quantum Fisher information; S62. Fusion of user feedback and promotion effects: With the help of quantum state evolution mechanism, the quantum master equation is used to fuse user feedback and promotion effects, and the knowledge graph is dynamically updated in real time, stably, and adaptively to feedback. The formula is: ； Where, For knowledge graph over time The evolution rate of The knowledge graph Lyapunov operator acts on the time The knowledge graph state at is the updated intensity coefficient, The quantum update unitary operator that adjusts the internal associations of the knowledge graph based on user behavior, To seek deviation operation; S63. Dynamically optimize algorithm parameters based on evaluation indicators, use quantum meta-learning to quickly adapt to new tasks of smart library reading promotion, and use quantum regularization to prevent overfitting to ensure the adaptability of promotion effect and knowledge graph quality. The formula is: ； Where, is the optimal algorithm parameter obtained after optimization, To cover various tasks in the promotion of reading in smart libraries Based on distribution Take expectations, Update parameters for measuring meta-learning The loss function of the performance error of the algorithm in the reading promotion task is: is the regularization strength, A quantum regularization term is used to prevent overfitting in the optimization of reading promotion parameters. S64. Process the knowledge graph from the perspectives of mining knowledge causal relationships, aligning multimodal information, and verifying strategy stability, strengthen the dynamic update capability of the knowledge graph, the multimodal fusion effect, and the strategy robustness, and continue to deepen the deep integration of quantum technology and library scenarios.

8. The method for promoting reading in a smart library based on knowledge graph according to claim 7 is characterized in that: In step S63, the meta-learning update parameters are used to quickly adapt to the new task of reading promotion. The formula is: ； Where, To control the learning rate of the meta-learning update during the parameter update step size, is the loss function Meta-learning parameters gradient.

9. The method for promoting reading in a smart library based on knowledge graph according to claim 7 is characterized in that: In step S64, processing the knowledge graph includes the following steps: S641. Distinguish the causal and false correlations between promotion strategies and user behaviors. Use quantum causal inference to identify the causal correlations between promotion strategies and user behaviors, and avoid false correlations interfering with updates. The formula is: ； Where, is the quantum causal coefficient, A quantum causal unitary operator that encodes causal propagation from strategy to behavior, To measure the quantum mutual information between strategy and behavior, To develop a reading promotion strategy for smart libraries, For user behavior; S642. Through quantum state superposition, align user text, space, and time series multimodal feedback data, resolve modal conflicts, and accurately update the knowledge graph. The formula is: ； Where, To uniformly update the multimodal superposition state of the knowledge graph, The modal weights reflect the degree of influence of different modes when superimposed, To verify the multimodal feedback state and knowledge graph subspace Quantum exchange test of alignment; S643. Use quantum random walks to simulate extreme scenarios and verify the robustness of the reading promotion strategy. This prevents the promotion strategy from being only adapted to conventional scenarios and ensures that the promotion strategy operates effectively in unpopular and sudden scenarios. The formula is: ； Where, is the robustness score for judging the stability of the strategy, Reading promotion strategies for smart libraries The adjoint operator of A quantum robust operator that imposes extreme scenario interference on the generalization strategy, is the worst-case density matrix, To perform trace operation.

Images