Material information semantic mapping control method based on natural language processing

Abstract

The invention discloses a material information semantic mapping control method based on natural language processing, and particularly relates to the field of data analysis, comprising material information acquisition, data analysis, semantic multivariate fusion potential excitation analysis, semantic toughness shaping and coevolution evaluation, semantic ecological adaptability expansion analysis and semantic mapping intelligent regulation decision. According to the method, multi-dimensional data of material information is comprehensively collected and analyzed, the semantic multivariate fusion potential is stimulated, a semantic structure with toughness is molded, the ecological adaptability of the semantic structure in different application scenes is evaluated, intelligent regulation and control of semantic mapping are achieved by means of an intelligent algorithm, the mapping effect is optimized, and the method is suitable for popularization and application. The accuracy and adaptability of material information semantics are improved, and powerful support is provided for efficient utilization and intelligent processing of material information.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and more specifically, to a material information semantic mapping control method based on natural language processing. Background Art

[0002] In the existing technology, the processing of material information mainly relies on automated information collection systems and database technology. The system will first automatically collect a large amount of material information from channels such as scientific research literature, experimental reports and patent databases, and organize and classify this information and store it in the database. When a specific material needs to be analyzed, researchers can use specific search conditions to quickly extract relevant material information from the database for subsequent analysis and research.

[0003] However, although the existing technology has improved the efficiency of material information processing to a certain extent, it still has some obvious shortcomings. First, the existing technology mainly relies on keyword matching for information retrieval, which makes it difficult to accurately capture the intrinsic connections and potential value between material information. Second, when processing large-scale, complex and diverse material information, the efficiency and accuracy of the existing technology still need to be improved. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a material information semantic mapping control method based on natural language processing, which solves the problems raised in the above-mentioned background technology through the following scheme.

[0005] To achieve the above object, the present invention provides the following technical solution: a material information semantic mapping control method based on natural language processing, comprising: Step 1: Material information collection: Collect semantic deep fusion data, semantic innovation and expansion data, semantic stability assessment data, and semantic application adaptation data of target material information; Step 2: Data analysis: used to analyze the data collected in step 1 and calculate the evaluation value of each set of data, including the semantic deep fusion evaluation model, the semantic innovation and expansion evaluation model, the semantic stability evaluation model, and the semantic application adaptation evaluation model; Step 3: Semantic multi-dimensional fusion potential stimulation analysis: used to establish a semantic fusion innovation potential index calculation model, used to analyze the basic ability of material information semantics to generate unique value in the initial stage of multi-dimensional fusion and expansion; Step 4: Semantic resilience shaping and co-evolution assessment: used to establish a semantic fusion innovation-stability co-evolution coefficient calculation model to observe how the fusion innovation potential analyzed in step 3 interacts with stability when facing challenges related to semantic stability, shaping a semantic structure that is resilient and capable of co-evolution; Step 5: Semantic ecological adaptability expansion analysis: used to establish a semantic ecological adaptability expansion coefficient calculation model. Based on steps 3 and 4, the ecological adaptability of semantics in different application scenarios is considered; Step 6: Semantic mapping intelligent control decision: According to the analysis results of step 5, the semantic mapping strategy is dynamically adjusted with the help of intelligent algorithms to optimize the semantic mapping effect of material information.

[0006] Preferably, the semantic deep fusion data include cross-modal semantic synergy coefficient, semantic hierarchical nesting score, semantic knowledge graph fusion degree and semantic metaphor association strength; the semantic innovation and expansion data include semantic concept generation rate, semantic logic reconstruction index, semantic association expansion width and semantic context migration adaptability; the semantic stability evaluation data include semantic fluctuation variance value, semantic ambiguity resolution rate, semantic noise tolerance coefficient and semantic redundancy elimination ratio; the semantic application adaptation data include semantic task matching degree, semantic process embedding fit, semantic decision support effectiveness and semantic cross-domain transplantation adaptability.

[0007] Preferably, the method for collecting the cross-modal semantic synergy coefficient is specifically as follows: first, the text and other modal data are respectively mapped to a common semantic space through a cross-modal feature extraction method, and then the covariance between the semantic feature vectors of different modalities is calculated, and then divided by the product of their respective standard deviations to obtain the cross-modal semantic synergy coefficient. The method for collecting the semantic hierarchical nesting score is specifically as follows: using syntax trees and semantic role labeling technology to analyze the hierarchical structure of semantic components in the material information text, assigning scores to reasonably nested semantic hierarchical pairs, giving 1 point for a complete match to the set ideal nesting pattern, and 0.5 points for a partial match, summarizing the scores of all hierarchical pairs and averaging them to obtain the semantic hierarchical nesting score. The specific method for collecting the fusion degree of the semantic knowledge graph is as follows: construct a knowledge graph in the material field, extract the semantic elements from the collected text semantic information and fuse them with the corresponding nodes and edges in the knowledge graph, and count the proportion of the number of successfully fused semantic elements to the total number of fused semantic elements in the text, which is used as the fusion degree of the semantic knowledge graph. The specific method for collecting the semantic metaphor association strength is as follows: through the metaphor recognition algorithm, identify the sentences in the material information text that use metaphors to describe the material, and then analyze the distance between the metaphor and the subject in the semantic feature space through the word vector method, and subtract the normalized value of the distance from 1 to obtain the semantic metaphor association strength.

[0008] Preferably, the method for collecting the semantic concept birth rate is specifically as follows: using text mining technology to analyze a large amount of the latest material information texts, combining professional field dictionaries and boundary entropy methods based on statistical language models, identifying completely new concepts that did not exist in the previous material semantic system, and counting the proportion of their number to the total number of concepts, which is the semantic concept birth rate. The method for collecting the semantic logic reconstruction index is specifically as follows: using a rule-based logic analyzer combined with a deep learning semantic parsing model to analyze the logical expressions of material causal relationships, function and structure associations in the material information text, and comparing them with traditional, established material semantic logic structures, counting the proportion of the number of sentences in the text that reconstructed the original logical relationship to the total number of logic-related sentences, and obtaining the semantic logic reconstruction index. The specific method for collecting the width of semantic association expansion is as follows: starting from the core concept of the material, calculating the semantic similarity of word vectors and finding the path from the core concept to other related concepts in the semantic network based on graph algorithms, counting the number of concepts of different fields and natures that can be associated within a certain text corpus, comparing it with the benchmark value of the number of traditional associated concepts, and calculating the expansion ratio as an indicator of the width of semantic association expansion. The specific method for collecting the adaptability of semantic context transfer is as follows: placing the material information text in different simulated contexts, and evaluating whether the text semantics can still be reasonably expressed, understood and coherent in these new contexts through the pre-trained language model, and counting the proportion of effective adaptation of semantics in different new contexts, that is, the adaptability of semantic context transfer.

[0009] Preferably, the method for collecting the variance value of the semantic fluctuation is specifically as follows: at different time points, the same key semantic dimension is quantified and represented, and the variance of these numerical sequences is calculated. The method for collecting the semantic ambiguity resolution rate is specifically as follows: material information texts from different backgrounds are collected, and for material performance index words with multiple interpretations, a Bayesian model is used in combination with semantic role analysis to count the proportion of the number of cases where ambiguity is successfully resolved to the total number of ambiguous cases, so as to obtain the semantic ambiguity resolution rate. The method for collecting the semantic noise tolerance coefficient is specifically as follows: different degrees of noise are artificially added to the material information text, and then a semantic understanding model based on deep learning is used to determine the proportion of the text that can still accurately convey the original semantics after the noise is added, so as to obtain the semantic noise tolerance coefficient. The method for collecting the semantic redundancy elimination ratio is specifically as follows: the semantic redundancy in the material information text is analyzed, and the importance is determined by a text compression algorithm based on the contribution of semantics to the expression of the core information of the material, and the proportion of the reduction in the amount of semantic information in the text after removing the redundancy is calculated, i.e., the semantic redundancy elimination ratio.

[0010] Preferably, the method for collecting the semantic task matching degree is as follows: for different practical application tasks related to materials, the semantics of the material information text are matched and compared with the key semantic elements required for each task, and the proportion of semantic intersection in the total semantic set is measured by set theory related algorithms combined with semantic similarity to obtain the semantic task matching degree. The method for collecting the semantic process embedding fit is as follows: according to the established process of material processing and application, by constructing a process semantic model, the degree of fit between the text semantics and the process semantics is calculated. The method for collecting the effectiveness of semantic decision support is as follows: in a decision-making scenario involving materials, the text semantics containing material information is used as input, and by combining expert systems and machine learning decision algorithms, the degree of fit between the decisions made based on the semantics and the actual optimal decisions is observed, and the effectiveness of semantic decision support is measured by the proportion of the number of fits in the total number of decisions. The method for collecting the adaptability of semantic cross-domain transplantation is as follows: trying to transplant the semantics of the material information text to other fields, and through representative texts in the target field combined with semantic similarity and logical coherence quantitative analysis methods, the proportion of semantics that can be applied after cross-domain transplantation is statistically analyzed to obtain the adaptability of semantic cross-domain transplantation.

[0011] Preferably, the semantic deep fusion evaluation model is used to evaluate the semantic deep fusion data, which is specifically expressed as: , SD represents the semantic deep fusion evaluation value, C represents the cross-modal semantic synergy coefficient, L represents the semantic hierarchical nesting score, G represents the semantic knowledge graph fusion degree, and I represents the semantic metaphor association strength.

[0012] Preferably, the semantic innovation and extension evaluation model is used to evaluate the semantic innovation and extension data, which is specifically expressed as follows: , SI represents the semantic innovation and expansion assessment value, R represents the semantic concept generation rate, E represents the semantic logic reconstruction index, W represents the semantic association expansion width, and M represents the semantic context transfer adaptability.

[0013] Preferably, the semantic stability evaluation model is used to evaluate the semantic stability evaluation data, which is specifically expressed as: , SS represents the semantic stability assessment value, T represents the temporal stability, D represents the semantic ambiguity resolution rate, V represents the semantic fluctuation variance value, N represents the semantic noise tolerance coefficient, and R represents the semantic redundancy elimination ratio.

[0014] Preferably, the semantic application adaptation evaluation model is used to evaluate the semantic application adaptation data, which is specifically expressed as: , SA represents the semantic application adaptation assessment value, S represents the semantic task matching degree, F represents the semantic process embedding fit, A represents the semantic decision support effectiveness, and P represents the semantic cross-domain transplantation adaptability.

[0015] Preferably, the semantic fusion innovation potential index calculation model is specifically expressed as: , P1 represents the semantic fusion innovation potential index, α represents the semantic deep fusion regulation index, β represents the semantic innovation expansion regulation index, and ε=0.001.

[0016] Preferably, the semantic fusion innovation-stability co-evolution coefficient calculation model is specifically expressed as: , P2 represents the semantic fusion innovation-stability co-evolution coefficient, and γ represents the semantic stability adjustment coefficient.

[0017] Preferably, the semantic ecological adaptability expansion coefficient calculation model is specifically expressed as: , P3 represents the semantic ecological adaptability expansion coefficient, and δ represents the semantic application adaptation adjustment parameter.

[0018] Preferably, step 6 inputs P3 into an intelligent decision-making model based on reinforcement learning, specifically adopting a deep Q network model, with P3 and historical data as state inputs, and different semantic mapping control actions as optional behaviors. The model is trained through simulation, and the training data is constructed based on past material information semantic mapping cases and corresponding control feedback results to learn the optimal control actions under different P3 value ranges.

[0019] Preferably, in step 6, when the decision indication of the model output is P3≥0.7, the "radical expansion" strategy is adopted, which means that on the basis of maintaining the existing semantic mapping state, new fusion modes, innovative concepts and application scenarios are actively explored, which is achieved by increasing the types of modalities of multimodal fusion and encouraging cross-domain semantic associations. When 0.3<P3<0.7, the "robust optimization" strategy is executed to optimize the relatively weak links in steps 4 and 5, which is specifically determined by backtracking the data and calculation process of each stage. When P3≤0.3, the "reconstruction and reshaping" strategy is implemented, which requires a comprehensive re-examination of the construction, fusion and application adaptation of the material information semantics, involving the replacement of the semantic analysis model, the re-planning of the semantic knowledge graph structure, and the redefinition of semantic elements based on actual application pain points.

[0020] Technical effects and advantages of the present invention: The present invention comprehensively collects data of multiple dimensions of material information, including semantic deep fusion data, semantic innovation and expansion data, semantic stability evaluation data and semantic application adaptation data, and ensures the accuracy and reliability of the data through specific data collection methods; conducts in-depth analysis of the data collected in step 1, and calculates the evaluation value of each group of data by constructing different evaluation models. These evaluation models comprehensively consider multiple factors and can comprehensively reflect the performance of material information in different dimensions, providing strong data support for subsequent steps such as semantic multi-dimensional fusion potential stimulation analysis, semantic resilience shaping and co-evolution evaluation, and providing a rich and high-quality information foundation for subsequent data analysis and semantic mapping control; establishes a semantic fusion innovation potential index calculation model to analyze the basic ability of material information semantics to generate unique value in the initial stage of multi-dimensional fusion and expansion. Through this model, the potential of semantics in multi-dimensional fusion can be quantitatively evaluated, which provides a clear direction for subsequent development and optimization, helps to discover new fusion models and innovative concepts, and promotes the further development of material information semantics; establishes a semantic fusion innovation-stability co-evolution coefficient calculator A calculation model is established to observe how the potential for fusion innovation interacts with stability when facing challenges related to semantic stability. The model can evaluate the performance of semantic structures in co-evolution and the impact of stability on the overall semantic structure, which helps to shape a resilient and co-evolving semantic structure and improve the adaptability and vitality of material information semantics. A calculation model for the semantic ecological adaptability expansion coefficient is established. On the basis of fusion innovation and stability shaping, the ecological adaptability of semantics in different application scenarios is considered. Through this model, the adaptive expansion of semantics in the application ecology can be quantitatively evaluated, which provides strong support for subsequent application development and optimization, helps to improve the adaptability and expansion effect of semantics in practical applications, and promotes the widespread application of material information semantics. According to the analysis results of step 5, the semantic mapping strategy is dynamically adjusted with the help of intelligent algorithms for control, which is used to optimize the material information semantic mapping effect. Through the intelligent decision-making model based on reinforcement learning, the optimal control actions under different conditions can be learned to realize the intelligent regulation of semantic mapping, which helps to improve the accuracy and efficiency of semantic mapping and promote the further development of material information semantic mapping technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is a schematic diagram of the overall structure of the present invention. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] refer to Figure 1 A material information semantic mapping control method based on natural language processing is shown, and the specific steps include: Step 1: Material information collection: Collect semantic deep fusion data, semantic innovation expansion data, semantic stability assessment data and semantic application adaptation data of target material information.

[0024] The semantic deep fusion data include cross-modal semantic synergy coefficient, semantic hierarchical nesting score, semantic knowledge graph fusion degree and semantic metaphor association strength; the semantic innovation and expansion data include semantic concept generation rate, semantic logic reconstruction index, semantic association expansion width and semantic context migration adaptability; the semantic stability assessment data include semantic fluctuation variance value, semantic ambiguity resolution rate, semantic noise tolerance coefficient and semantic redundancy elimination ratio; the semantic application adaptation data include semantic task matching degree, semantic process embedding fit, semantic decision support effectiveness and semantic cross-domain transplantation adaptability.

[0025] The method for collecting the cross-modal semantic synergy coefficient is specifically as follows: first, the text and other modal data are respectively mapped to a common semantic space through a cross-modal feature extraction method, and then the covariance between the semantic feature vectors of different modalities is calculated, and then divided by the product of their respective standard deviations to obtain the cross-modal semantic synergy coefficient. The method for collecting the semantic level nesting score is specifically as follows: the syntax tree and semantic role labeling technology are used to analyze the hierarchical structure of the semantic components in the material information text, and scores are assigned to reasonably nested semantic level pairs. A complete match with the set ideal nesting mode is scored as 1 point, and a partial match is scored as 0.5 points. The scores of all level pairs are summarized and averaged to obtain the semantic level nesting score. The specific method for collecting the fusion degree of the semantic knowledge graph is as follows: construct a knowledge graph in the material field, extract the semantic elements from the collected text semantic information and fuse them with the corresponding nodes and edges in the knowledge graph, and count the proportion of the number of successfully fused semantic elements to the total number of fused semantic elements in the text, which is used as the fusion degree of the semantic knowledge graph. The specific method for collecting the semantic metaphor association strength is as follows: through the metaphor recognition algorithm, identify the sentences in the material information text that use metaphors to describe the material, and then analyze the distance between the metaphor and the subject in the semantic feature space through word vectors, and subtract the normalized value of the distance from 1 to obtain the semantic metaphor association strength.

[0026] The method for collecting the semantic concept birth rate is specifically as follows: using text mining technology to analyze a large amount of the latest material information texts, combining professional field dictionaries and the boundary entropy method based on statistical language models, identifying completely new concepts that did not exist in the previous material semantic system, and counting the proportion of their number to the total number of concepts, which is the semantic concept birth rate. The method for collecting the semantic logic reconstruction index is specifically as follows: using a rule-based logic analyzer combined with a deep learning semantic parsing model to analyze the logical expressions of material causal relationships, function and structure associations in the material information text, and comparing them with the traditional and established material semantic logic structure, counting the proportion of the number of sentences in the text that reconstructed the original logical relationship to the total number of logic-related sentences, and obtaining the semantic logic reconstruction index. The specific method for collecting the width of semantic association expansion is as follows: starting from the core concept of the material, calculating the semantic similarity of word vectors and finding the path from the core concept to other related concepts in the semantic network based on graph algorithms, counting the number of concepts of different fields and natures that can be associated within a certain text corpus, comparing it with the benchmark value of the number of traditional related concepts, and calculating the expansion ratio as an indicator of the width of semantic association expansion. The specific method for collecting the adaptability of semantic context transfer is as follows: placing the material information text in different simulated contexts, and using the pre-trained language model to evaluate whether the text semantics can still be reasonably expressed, understood and coherent in these new contexts, and counting the proportion of effective adaptation of semantics in different new contexts, that is, semantic context transfer adaptability.

[0027] The method for collecting the variance value of semantic fluctuation is specifically as follows: at different time points, the same key semantic dimension is quantified and represented, and the variance of these numerical sequences is calculated. The method for collecting the semantic ambiguity resolution rate is specifically as follows: material information texts from different backgrounds are collected, and for material performance indicator words with multiple interpretations, the Bayesian model is combined with semantic role analysis to count the proportion of the number of cases where ambiguity is successfully resolved to the total number of ambiguous cases, so as to obtain the semantic ambiguity resolution rate. The method for collecting the semantic noise tolerance coefficient is specifically as follows: different degrees of noise are artificially added to the material information text, and then the semantic understanding model based on deep learning is used to judge the proportion of the text that can still accurately convey the original semantics after the noise is added, so as to obtain the semantic noise tolerance coefficient. The method for collecting the semantic redundancy elimination ratio is specifically as follows: the semantic redundancy in the material information text is analyzed, and the importance is judged according to the contribution of semantics to the expression of the core information of the material through a text compression algorithm, and the proportion of the reduction in the amount of semantic information in the text after removing the redundancy is calculated, that is, the semantic redundancy elimination ratio.

[0028] The method for collecting the semantic task matching degree is specifically as follows: for different practical application tasks related to materials, the semantics of the material information text are matched and compared with the key semantic elements required for each task, and the semantic task matching degree is obtained by measuring the proportion of semantic intersections in the total semantic set through set theory-related algorithms combined with semantic similarity. The method for collecting the semantic process embedding fit is specifically as follows: according to the established process of material processing and application, the degree of fit between the text semantics and the process semantics is calculated by constructing a process semantic model. The method for collecting the effectiveness of semantic decision support is specifically as follows: in a decision-making scenario involving materials, the text semantics containing material information is used as input, and by combining expert systems and machine learning decision algorithms, the degree of fit between the decisions made based on the semantics and the actual optimal decisions is observed, and the effectiveness of semantic decision support is measured by the proportion of the number of fits in the total number of decisions. The method for collecting the adaptability of semantic cross-domain transplantation is specifically as follows: an attempt is made to transplant the semantics of the material information text to other fields, and through representative texts in the target field combined with semantic similarity and logical coherence quantitative analysis methods, the proportion of semantics that can be applied after cross-domain transplantation is statistically analyzed to obtain the adaptability of semantic cross-domain transplantation.

[0029] Step 2: Data analysis: used to analyze the data collected in step 1 and calculate the evaluation value of each set of data, including the semantic deep fusion evaluation model, the semantic innovation and expansion evaluation model, the semantic stability evaluation model and the semantic application adaptation evaluation model.

[0030] The semantic deep fusion evaluation model is used to evaluate the semantic deep fusion data, which is specifically expressed as follows: , SD represents the semantic deep fusion evaluation value, C represents the cross-modal semantic synergy coefficient, L represents the semantic hierarchical nesting score, G represents the semantic knowledge graph fusion degree, and I represents the semantic metaphor association strength.

[0031] In the semantic deep fusion evaluation model, ln(C+1) reflects the marginal diminishing effect of cross-modal collaboration, arctan(L / π) maps the hierarchical nesting score to a finite interval, tanh(G) is used to normalize the influence of the knowledge graph fusion degree, √(I^3+1) emphasizes the nonlinear growth of metaphor association, and the denominator e^((C+L) / (G+I)) is used as a regulating factor to balance the influence of each parameter.

[0032] The semantic innovation and extension evaluation model is used to evaluate the semantic innovation and extension data, which is specifically expressed as follows: , SI represents the semantic innovation and expansion assessment value, R represents the semantic concept generation rate, E represents the semantic logic reconstruction index, W represents the semantic association expansion width, and M represents the semantic context transfer adaptability.

[0033] In the semantic innovation and expansion evaluation model, (R×ln(E+1))^2 represents the synergistic effect of concept generation and logic reconstruction, sinh(W / 10) reflects the exponential growth characteristics of associative expansion, |cos(M×π)| introduces periodic fluctuation correction, and √(R×E×W×M) is used as the basic guarantee value.

[0034] The semantic stability evaluation model is used to evaluate the semantic stability evaluation data, and is specifically expressed as follows: , SS represents the semantic stability assessment value, T represents the temporal stability, D represents the semantic ambiguity resolution rate, V represents the semantic fluctuation variance value, N represents the semantic noise tolerance coefficient, and R represents the semantic redundancy elimination ratio.

[0035] In the semantic stability evaluation model, (T^2+D^2)^(1 / 3) calculates the comprehensive effect of stability and ambiguity resolution, exp(-V / N) reflects the negative impact of fluctuations on stability, sin(πR / 180) introduces the periodic effect of redundancy elimination, and log(V×N+1) is used as a normalization factor.

[0036] The semantic application adaptation evaluation model is used to evaluate the semantic application adaptation data, which is specifically expressed as follows: , SA represents the semantic application adaptation assessment value, S represents the semantic task matching degree, F represents the semantic process embedding fit, A represents the semantic decision support effectiveness, and P represents the semantic cross-domain transplantation adaptability.

[0037] In the semantic application adaptation evaluation model, the cubic mean of (S^3+F^3)^(1 / 3) reflects the basic adaptation capability, exp(-(AP)^2 / 10) evaluates the matching degree between decision support and transplantation adaptability, sinh((A+P) / 4) emphasizes the synergistic effect of decision and transplantation, and log(S×F+2) is used as a constraint to ensure that the evaluation value is within a reasonable range.

[0038] Step 3: Semantic multi-dimensional fusion potential stimulation analysis: used to establish a semantic fusion innovation potential index calculation model to analyze the basic ability of material information semantics to generate unique value in the initial stage of multi-dimensional fusion and expansion.

[0039] The semantic fusion innovation potential index calculation model is specifically expressed as: , P1 represents the semantic fusion innovation potential index, α represents the semantic deep fusion regulation index, β represents the semantic innovation expansion regulation index, and ε=0.001.

[0040] The ε is used to avoid the situation where the denominator is 0 and is a very small positive number.

[0041] The semantic fusion innovation potential index calculation model performs power operations on SD and SI and combines the difference processing of the denominator. It not only considers the individual strengths of the two, but also pays attention to the relative balance relationship between them, so as to measure the initial potential of semantic multi-fusion. The higher the P1 value, the stronger the potential for semantics to be explored from the perspective of fusion and innovation, providing a good start for subsequent development.

[0042] Step 4: Semantic resilience shaping and co-evolution assessment: used to establish a semantic fusion innovation-stability co-evolution coefficient calculation model to observe how the fusion innovation potential analyzed in step 3 interacts with stability when facing challenges related to semantic stability, and to shape a resilient and co-evolving semantic structure.

[0043] The semantic fusion innovation-stability co-evolution coefficient calculation model is specifically expressed as: , P2 represents the semantic fusion innovation-stability co-evolution coefficient, and γ represents the semantic stability adjustment coefficient.

[0044] The semantic fusion innovation-stability co-evolution coefficient calculation model multiplies P1 in step 3 by a function about SS, in which SS is exponentially operated and combined with the special construction of the denominator, 1 is subtracted from SS and a very small positive number ε is added to highlight the positive impact of good stability. At the same time, an exponential operation is performed with the logarithm of the sum of SD and SI as the base, which links stability with the potential for fusion innovation and takes into account the impact of stability on the overall semantic structure in co-evolution. The larger the P2 value, the better the semantics can shape resilient and co-evolutionary characteristics with the help of stability on the basis of fusion innovation.

[0045] Step 5: Semantic ecological adaptability expansion analysis: used to establish a semantic ecological adaptability expansion coefficient calculation model. Based on steps 3 and 4, the ecological adaptability of semantics in different application scenarios is considered.

[0046] The semantic ecological adaptability expansion coefficient calculation model is specifically expressed as: , P3 represents the semantic ecological adaptability expansion coefficient, and δ represents the semantic application adaptation adjustment parameter.

[0047] The semantic ecological adaptability expansion coefficient calculation model performs a power operation on SA and multiplies it by an expression combining a trigonometric function and an exponential function. The trigonometric function This makes SA present a nonlinear characteristic of first increasing and then decreasing when it changes in the range of 0 to 1, which can better characterize the influence of application adaptation at different stages. The exponential function part further adjusts the curve shape to measure the adaptive expansion of semantics in the application ecology as a whole. The higher the P3 value, the stronger the adaptability of semantics in the actual application ecology after experiencing integrated innovation and stability shaping, and the better the expansion effect.

[0048] Step 6: Semantic mapping intelligent control decision: According to the analysis results of step 5, the semantic mapping strategy is dynamically adjusted with the help of intelligent algorithms to optimize the semantic mapping effect of material information.

[0049] The step 6 inputs P3 into the intelligent decision-making model based on reinforcement learning, specifically adopting a deep Q network model, with P3 and historical data as state inputs, and different semantic mapping control actions as optional behaviors. The model is trained through simulation, and the training data is constructed based on past material information semantic mapping cases and corresponding control feedback results to learn the optimal control actions under different P3 value ranges.

[0050] In step 6, when the decision indication of the model output is P3≥0.7, the "radical expansion" strategy is adopted, which means that on the basis of maintaining the existing semantic mapping state, new fusion modes, innovative concepts and application scenarios are actively explored, which is achieved by increasing the types of multimodal fusion modes and encouraging cross-domain semantic associations. When 0.3<P3<0.7, the "robust optimization" strategy is executed to optimize the relatively weak links in steps 4 and 5, which is specifically determined by tracing back the data and calculation process of each stage. When P3≤0.3, the "reconstruction and reshaping" strategy is implemented, which requires a comprehensive re-examination of the construction, fusion and application adaptation of the material information semantics, involving the replacement of the semantic analysis model, the re-planning of the semantic knowledge graph structure, and the redefinition of semantic elements based on actual application pain points.

[0051] The present invention comprehensively collects data of multiple dimensions of material information, including semantic deep fusion data, semantic innovation and expansion data, semantic stability evaluation data and semantic application adaptation data, and ensures the accuracy and reliability of the data through specific data collection methods; conducts in-depth analysis of the data collected in step 1, and calculates the evaluation value of each group of data by constructing different evaluation models. These evaluation models comprehensively consider multiple factors and can comprehensively reflect the performance of material information in different dimensions, providing strong data support for subsequent steps such as semantic multi-dimensional fusion potential stimulation analysis, semantic resilience shaping and co-evolution evaluation, and providing a rich and high-quality information foundation for subsequent data analysis and semantic mapping control; establishes a semantic fusion innovation potential index calculation model to analyze the basic ability of material information semantics to generate unique value in the initial stage of multi-dimensional fusion and expansion. Through this model, the potential of semantics in multi-dimensional fusion can be quantitatively evaluated, which provides a clear direction for subsequent development and optimization, helps to discover new fusion models and innovative concepts, and promotes the further development of material information semantics; establishes a semantic fusion innovation-stability co-evolution coefficient calculator A calculation model is established to observe how the potential for fusion innovation interacts with stability when facing challenges related to semantic stability. The model can evaluate the performance of semantic structures in co-evolution and the impact of stability on the overall semantic structure, which helps to shape a resilient and co-evolving semantic structure and improve the adaptability and vitality of material information semantics. A calculation model for the semantic ecological adaptability expansion coefficient is established. On the basis of fusion innovation and stability shaping, the ecological adaptability of semantics in different application scenarios is considered. Through this model, the adaptive expansion of semantics in the application ecology can be quantitatively evaluated, which provides strong support for subsequent application development and optimization, helps to improve the adaptability and expansion effect of semantics in practical applications, and promotes the widespread application of material information semantics. According to the analysis results of step 5, the semantic mapping strategy is dynamically adjusted with the help of intelligent algorithms for control, which is used to optimize the material information semantic mapping effect. Through the intelligent decision-making model based on reinforcement learning, the optimal control actions under different conditions can be learned to realize the intelligent regulation of semantic mapping, which helps to improve the accuracy and efficiency of semantic mapping and promote the further development of material information semantic mapping technology.

[0052] Secondly: In the drawings of the embodiments disclosed in the present invention, only the structures related to the embodiments disclosed in the present invention are involved, and other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of the present invention can be combined with each other; Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A material information semantic mapping control method based on natural language processing, characterized in that: include: Step 1: Material information collection: Collect semantic deep fusion data, semantic innovation and expansion data, semantic stability assessment data, and semantic application adaptation data of target material information; Step 2: Data analysis: used to analyze the data collected in step 1 and calculate the evaluation value of each set of data, including the semantic deep fusion evaluation model, the semantic innovation and expansion evaluation model, the semantic stability evaluation model, and the semantic application adaptation evaluation model; Step 3: Semantic multi-dimensional fusion potential stimulation analysis: used to establish a semantic fusion innovation potential index calculation model, used to analyze the basic ability of material information semantics to generate unique value in the initial stage of multi-dimensional fusion and expansion; Step 4: Semantic resilience shaping and co-evolution assessment: used to establish a semantic fusion innovation-stability co-evolution coefficient calculation model to observe how the fusion innovation potential analyzed in step 3 interacts with stability when facing challenges related to semantic stability, shaping a semantic structure that is resilient and capable of co-evolution; Step 5: Semantic ecological adaptability expansion analysis: used to establish a semantic ecological adaptability expansion coefficient calculation model. Based on steps 3 and 4, the ecological adaptability of semantics in different application scenarios is considered; Step 6: Semantic mapping intelligent control decision: According to the analysis results of step 5, the semantic mapping strategy is dynamically adjusted with the help of intelligent algorithms to optimize the semantic mapping effect of material information.

2. The material information semantic mapping control method based on natural language processing according to claim 1 is characterized by: The semantic deep fusion data include cross-modal semantic synergy coefficient, semantic hierarchical nesting score, semantic knowledge graph fusion degree and semantic metaphor association strength; the semantic innovation and expansion data include semantic concept generation rate, semantic logic reconstruction index, semantic association expansion width and semantic context migration adaptability; the semantic stability assessment data include semantic fluctuation variance value, semantic ambiguity resolution rate, semantic noise tolerance coefficient and semantic redundancy elimination ratio; the semantic application adaptation data include semantic task matching degree, semantic process embedding fit, semantic decision support effectiveness and semantic cross-domain transplantation adaptability.

3. The material information semantic mapping control method based on natural language processing according to claim 1 is characterized in that: The semantic deep fusion evaluation model is used to evaluate the semantic deep fusion data, which is specifically expressed as follows: , SD represents the semantic deep fusion evaluation value, C represents the cross-modal semantic synergy coefficient, L represents the semantic hierarchical nesting score, G represents the semantic knowledge graph fusion degree, and I represents the semantic metaphor association strength.

4. The material information semantic mapping control method based on natural language processing according to claim 1 is characterized in that: The semantic innovation and extension evaluation model is used to evaluate the semantic innovation and extension data, which is specifically expressed as follows: , SI represents the semantic innovation and expansion evaluation value, R represents the semantic concept generation rate, E represents the semantic logic reconstruction index, W represents the semantic association expansion width, and M represents the semantic context transfer adaptability.

5. The material information semantic mapping control method based on natural language processing according to claim 1 is characterized in that: The semantic stability evaluation model is used to evaluate the semantic stability evaluation data, and is specifically expressed as follows: , SS represents the semantic stability assessment value, T represents the temporal stability, D represents the semantic ambiguity resolution rate, V represents the semantic fluctuation variance value, N represents the semantic noise tolerance coefficient, and R represents the semantic redundancy elimination ratio.

6. The material information semantic mapping control method based on natural language processing according to claim 1 is characterized by: The semantic application adaptation evaluation model is used to evaluate the semantic application adaptation data, which is specifically expressed as follows: , SA represents the semantic application adaptation assessment value, S represents the semantic task matching degree, F represents the semantic process embedding fit, A represents the semantic decision support effectiveness, and P represents the semantic cross-domain transplantation adaptability.

7. The material information semantic mapping control method based on natural language processing according to claim 1 is characterized by: The semantic fusion innovation potential index calculation model is specifically expressed as: , P1 represents the semantic fusion innovation potential index, α represents the semantic deep fusion regulation index, β represents the semantic innovation expansion regulation index, and ε=0.

001.

8. The material information semantic mapping control method based on natural language processing according to claim 1 is characterized by: The semantic fusion innovation-stability co-evolution coefficient calculation model is specifically expressed as: , P2 represents the semantic fusion innovation-stability co-evolution coefficient, and γ represents the semantic stability adjustment coefficient.

9. The material information semantic mapping control method based on natural language processing according to claim 1, characterized in that: The semantic ecological adaptability expansion coefficient calculation model is specifically expressed as: , P3 represents the semantic ecological adaptability expansion coefficient, and δ represents the semantic application adaptation adjustment parameter.

Images