Knowledge graph construction method and device based on material reserve information

Abstract

The invention discloses a knowledge graph construction method and device based on material reserve information. The method comprises the following steps: acquiring material reserve information and material classification information; preprocessing the material reserve information to obtain preprocessed material reserve information; and processing the preprocessed material reserve information and the material classification information to obtain material reserve knowledge graph information. Visibly, by combining the category, similarity, semantic information and the like of the material reserve information, the material reserve knowledge graph information of the material reserve information can be accurately and efficiently obtained, association analysis of the materials is facilitated, and therefore powerful support is provided for reasonable configuration and emergency management of the materials, and the material reserve knowledge graph information of the material reserve information can be accurately and efficiently obtained. And the optimization efficiency and the emergency response efficiency of material reserve and supply chain management are improved.

Description

Technical Field

[0001] The present invention relates to the field of material management and information processing technology, and in particular to a method and device for constructing a knowledge graph based on material reserve information. Background Art

[0002] Modern material management faces challenges from increasing supply chain complexity and growing demand for emergency reserves, particularly in areas such as data fragmentation, insufficient intelligent analysis, slow response times, and difficulty establishing material linkages. Current material management methods rely primarily on manual statistics, fixed rule matching, and simple database retrieval. This results in fragmented material information, inaccurate decision-making, and untimely responses, easily leading to overstocking or shortages. Furthermore, the functions, uses, and substitution relationships between materials are difficult to automatically identify and analyze, hindering the optimal allocation of materials and the efficiency of emergency deployment. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a method and device for constructing a knowledge graph based on material reserve information. By combining the category, similarity and semantic information of the material reserve information, the material reserve knowledge graph information of the material reserve information can be obtained accurately and efficiently, which is conducive to realizing the correlation analysis of materials, thereby providing strong support for the reasonable allocation of materials and emergency management, and thus improving the optimization efficiency and emergency response efficiency of material reserves and supply chain management.

[0004] In order to solve the above technical problems, the first aspect of the embodiments of the present invention discloses a method for constructing a knowledge graph based on material reserve information, the method comprising:

[0005] S1, obtain material reserve information and material classification information;

[0006] S2, preprocessing the material reserve information to obtain preprocessed material reserve information;

[0007] S3: Process the pre-processed material reserve information and the material classification information to obtain material reserve knowledge graph information.

[0008] As an optional implementation manner, in the first aspect of the embodiment of the present invention, preprocessing the material reserve information to obtain preprocessed material reserve information includes:

[0009] S21, removing stop words from the material reserve information to obtain first material reserve preprocessing information;

[0010] S22, removing special characters from the first material reserve preprocessing information to obtain second material reserve preprocessing information;

[0011] S23, performing formatting uniform processing on the second material reserve preprocessing information to obtain third material reserve preprocessing information;

[0012] S24, performing synonym conversion processing on the third material reserve preprocessing information to obtain preprocessed material reserve information.

[0013] As an optional implementation manner, in the first aspect of the embodiment of the present invention, the processing of the pre-processed material reserve information and the material classification information to obtain material reserve knowledge graph information includes:

[0014] S31, obtaining a material reserve training data set;

[0015] S32, processing the material reserve training data set and the pre-processed material reserve information to obtain material reserve relationship information;

[0016] S33: Process the material reserve relationship information and the material classification information to obtain material reserve knowledge graph information.

[0017] As an optional implementation manner, in the first aspect of the embodiment of the present invention, the processing of the material reserve training dataset and the pre-processed material reserve information to obtain the material reserve relationship information includes:

[0018] S321, processing the pre-processed material reserve information to obtain unprocessed material reserve information;

[0019] S322, processing the to-be-processed material reserve information and the material reserve training data set to obtain material reserve sequence information;

[0020] S323: Process the material reserve sequence information to obtain material reserve relationship information.

[0021] As an optional implementation manner, in the first aspect of the embodiment of the present invention, the processing of the pre-processed material reserve information to obtain the unprocessed material reserve information includes:

[0022] S3211, performing feature word extraction processing on the pre-processed material reserve information to obtain material reserve feature word information;

[0023] S3212: Process the material reserve feature word information and the pre-processed material reserve information to obtain a material reserve feature word vector set;

[0024] S3213, using a material reserve semantic calculation model, calculating and processing the material reserve feature word vector set to obtain material reserve information to be processed;

[0025] The material reserve semantic computing model is:

[0026]

[0027] Where, V is the reserve information of materials to be processed, V i1 is the i1th material reserve semantic vector in the material reserve information to be processed, CXL is the material reserve feature word vector set, CXL i1 and CXL j1 are the i1th and j1th material reserve feature word vectors in the material reserve feature word vector set, N1 is the number of material reserve feature word vectors in the material reserve feature word vector set, δ1 is the first weight parameter, ∈ is the material reserve semantic deviation vector information, ∈ i1 It is the i1th material reserve semantic deviation vector in the material reserve semantic deviation vector information.

[0028] As an optional implementation manner, in the first aspect of the embodiment of the present invention, the processing of the material reserve sequence information to obtain the material reserve relationship information includes:

[0029] S3231, using a material reserve similarity calculation model, calculating and processing the material reserve sequence information to obtain material reserve similarity information;

[0030] The material reserve similarity calculation model is:

[0031]

[0032] Where, XSD is the similarity information of the material reserves, XSD i3,j3 is the material reserve similarity value between the i3th material and the j3th material in the material reserve similarity information, WZCB i3 and WZCB j3 are the i3th and j3th material reserve sequences in the material reserve sequence information, respectively, |·| represents the modulus of the vector, and θ1 represents the first similarity offset parameter;

[0033] S3232, performing category matching processing on the material reserve similarity information to obtain material reserve category information;

[0034] S3233: Calculate and process the material reserve category information to obtain material reserve relationship information.

[0035] As an optional implementation manner, in the first aspect of the embodiment of the present invention, the processing of the material reserve relationship information and the material classification information to obtain the material reserve knowledge graph information includes:

[0036] S331, processing the material reserve relationship information and the material classification information to obtain material classification sequence information;

[0037] S332, fusing the material classification sequence information and the material reserve relationship information to obtain a material classification matrix;

[0038] S333: Integrate the material classification matrix to obtain material reserve knowledge graph information.

[0039] A second aspect of an embodiment of the present invention discloses a device for constructing a knowledge graph based on material reserve information, the device comprising:

[0040] Acquisition module, used to obtain material reserve information and material classification information;

[0041] A preprocessing module, configured to preprocess the material reserve information to obtain preprocessed material reserve information;

[0042] The calculation module is used to process the pre-processed material reserve information and the material classification information to obtain material reserve knowledge graph information.

[0043] A third aspect of an embodiment of the present invention discloses another device for constructing a knowledge graph based on material reserve information, the device comprising:

[0044] processor;

[0045] a memory coupled to the processor and storing executable program code;

[0046] The processor calls the executable program code stored in the memory to execute some or all steps of the method for constructing a knowledge graph based on material reserve information disclosed in the first aspect of the embodiment of the present invention.

[0047] The fourth aspect of an embodiment of the present invention discloses a computer-readable storage medium, which stores computer instructions. When the computer instructions are called, they are used to execute part or all of the steps of the knowledge graph construction method based on material reserve information disclosed in the first aspect of an embodiment of the present invention.

[0048] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0049] In this embodiment of the present invention, material reserve information and material classification information are obtained; the material reserve information is preprocessed to obtain preprocessed material reserve information; and the preprocessed material reserve information and the material classification information are processed to obtain material reserve knowledge graph information. This embodiment, by combining the categories, similarities, and semantic information of the material reserve information, can accurately and efficiently obtain material reserve knowledge graph information for the material reserve information, facilitating the analysis of material associations, thereby providing strong support for the rational allocation of materials and emergency management, and thereby improving the optimization efficiency of material reserve and supply chain management and the efficiency of emergency response. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. 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 creative work.

[0051] Figure 1 A schematic diagram of a flow chart of a method for constructing a knowledge graph based on material reserve information disclosed in an embodiment of the present invention;

[0052] Figure 2 This is a schematic diagram of the structure of a device for constructing a knowledge graph based on material reserve information disclosed in an embodiment of the present invention;

[0053] Figure 3 This is a schematic structural diagram of another device for constructing a knowledge graph based on material reserve information disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0054] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of 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 making any creative efforts shall fall within the scope of protection of the present invention.

[0055] The terms "first," "second," and so on, in the description and claims of the present invention and the accompanying drawings are used to distinguish between different objects, not to describe a specific order. Furthermore, the terms "including," "having," and any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, apparatus, product, or device comprising a series of steps or elements is not limited to the listed steps or elements but may optionally include steps or elements not listed, or may optionally include other steps or elements inherent to the process, method, product, or device.

[0056] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present invention. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute a separate or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0057] The present invention discloses a method and device for constructing a knowledge graph based on material reserve information. By combining the categories, similarities, and semantic information of material reserve information, this method can accurately and efficiently obtain material reserve knowledge graph information, facilitating the analysis of material associations, thereby providing strong support for the rational allocation of materials and emergency management, thereby improving the optimization efficiency of material reserve and supply chain management and the efficiency of emergency response. These are described in detail below.

[0058] Example 1

[0059] See also Figure 1 , Figure 1 This is a flow chart of a method for constructing a knowledge graph based on material reserve information disclosed in an embodiment of the present invention. Figure 1 The method for constructing a knowledge graph based on material reserve information is applied to a device for constructing a knowledge graph based on material reserve information, such as a local server or cloud server for optimizing management of a knowledge graph based on material reserve information, and is not limited in the embodiments of the present invention. Figure 1 As shown, the method for constructing a knowledge graph based on material reserve information may include the following operations:

[0060] S1, obtain material reserve information and material classification information;

[0061] It should be noted that material reserve information refers to all structured and unstructured data related to materials, covering the basic properties of materials, reserve status, supply chain information, etc.; material classification information is used to organize and classify materials in a hierarchical manner, ensuring that the storage, management, and call of different categories of materials are more standardized, and providing structured data support for the construction of knowledge graphs.

[0062] S2, preprocessing the material reserve information to obtain preprocessed material reserve information;

[0063] S3, processing the pre-processed material reserve information and material classification information to obtain material reserve knowledge graph information.

[0064] It should be noted that the material reserve knowledge graph information is based on the material reserve information, and the multi-dimensional structured representation information is constructed to represent the entities, attributes and relationships between materials, thereby providing strong support for the rational allocation of materials and emergency management, thereby improving the optimization efficiency of material reserves and supply chain management and the efficiency of emergency response.

[0065] It can be seen that the implementation of the knowledge graph construction method based on material reserve information described in the embodiment of the present invention is conducive to realizing the correlation analysis of materials, thereby providing strong support for the rational allocation of materials and emergency management, and further improving the optimization efficiency of material reserves and supply chain management and the emergency response efficiency.

[0066] In an optional embodiment, preprocessing the material reserve information to obtain preprocessed material reserve information includes:

[0067] S21, removing stop words from the material reserve information to obtain first material reserve preprocessing information;

[0068] It should be noted that the above-mentioned process of removing stop words can be performed using tools or algorithms such as SpaCy and NLTK, and the specific embodiments of the present invention are not limited thereto.

[0069] It should be noted that by removing stop words, such as "的", "是", "了", etc., key material information can be made more prominent, thereby improving the accuracy of subsequent feature extraction and classification.

[0070] S22, removing special characters from the first material reserve preprocessing information to obtain second material reserve preprocessing information;

[0071] It should be noted that the above-mentioned removal of special characters can be performed using Python's re.sub() method, which can batch clean up useless information such as punctuation marks, web links, etc. The specific processing method is not specifically limited in the embodiment of the present invention.

[0072] It should be noted that through the above processing, interference characters such as #, *, / , @, and % can be removed, thereby improving the normalization of the data and ensuring that the model will not be affected by irrelevant characters.

[0073] S23, performing formatting uniform processing on the second material reserve preprocessing information to obtain third material reserve preprocessing information;

[0074] It should be noted that the above formats are processed uniformly and can be processed using tools or algorithms such as NLTK and SpaCy. The specific embodiments of the present invention do not limit this.

[0075] It should be noted that unified formatting ensures that there will be no information inconsistency due to differences in case or units when matching data.

[0076] S24 , performing synonym conversion processing on the third material reserve preprocessing information to obtain preprocessed material reserve information.

[0077] It should be noted that the above synonym conversion process can be performed using tools or algorithms such as NLTK and Word2Vec, and the embodiments of the present invention do not make specific limitations.

[0078] It should be noted that the above processing can ensure that data analysis and knowledge graph construction will not be mismatched due to different expressions.

[0079] It can be seen that the implementation of the knowledge graph construction method based on material reserve information described in the embodiment of the present invention is conducive to realizing the correlation analysis of materials, thereby providing strong support for the rational allocation of materials and emergency management, and further improving the optimization efficiency of material reserves and supply chain management and the emergency response efficiency.

[0080] In another optional embodiment, the pre-processed material reserve information and material classification information are processed to obtain material reserve knowledge graph information, including:

[0081] S31, obtaining a material reserve training data set;

[0082] It should be noted that the material reserve training dataset is a structured and cleaned data set that contains text, classification, relationship and attribute information related to material reserves, which is used to train the model to generate material reserve relationship information of material reserve information.

[0083] S32, processing the material reserve training data set and the pre-processed material reserve information to obtain material reserve relationship information;

[0084] S33, processing the material reserve relationship information and the material classification information to obtain material reserve knowledge graph information.

[0085] It can be seen that the implementation of the knowledge graph construction method based on material reserve information described in the embodiment of the present invention is conducive to realizing the correlation analysis of materials, thereby providing strong support for the rational allocation of materials and emergency management, and further improving the optimization efficiency of material reserves and supply chain management and the emergency response efficiency.

[0086] In yet another optional embodiment, the material reserve training data set and the pre-processed material reserve information are processed to obtain material reserve relationship information, including:

[0087] S321, processing the pre-processed material reserve information to obtain the material reserve information to be processed;

[0088] S322, processing the pending material reserve information and the material reserve training data set to obtain material reserve sequence information;

[0089] S323: Process the material reserve sequence information to obtain material reserve relationship information.

[0090] It can be seen that the implementation of the knowledge graph construction method based on material reserve information described in the embodiment of the present invention is conducive to realizing the correlation analysis of materials, thereby providing strong support for the rational allocation of materials and emergency management, and further improving the optimization efficiency of material reserves and supply chain management and the emergency response efficiency.

[0091] In an optional embodiment, the pre-processed material reserve information is processed to obtain the material reserve information to be processed, including:

[0092] S3211, extracting characteristic words from the pre-processed material reserve information to obtain material reserve characteristic word information;

[0093] It should be noted that the above-mentioned feature word extraction process can be performed using tools or algorithms such as TF-IDF, TextRank or BERT, and the specific embodiments of the present invention do not limit this.

[0094] It should be noted that the above processing can extract key feature words representing materials and reduce the interference of common words. At the same time, it can identify specific information such as material name, quantity, specifications, etc., to ensure that the multi-dimensional features of the materials can be extracted.

[0095] S3212, processing the material reserve feature word information and the pre-processed material reserve information to obtain a material reserve feature word vector set;

[0096] It should be noted that the above processing can be performed using tools or algorithms such as Word2Vec, GloVe or BERT, and the specific embodiments of the present invention are not limited thereto.

[0097] It should be noted that through the above processing, semantic vectors that can represent material feature words are generated, which is convenient for subsequent calculation of similarity, classification and other tasks. At the same time, a context-aware vector can be provided for each material feature word, thereby enhancing the model's ability to process complex sentences and polysemous words.

[0098] S3213, using the material reserve semantic calculation model, calculates and processes the material reserve feature word vector set to obtain the material reserve information to be processed;

[0099] Among them, the semantic computing model of material reserves is:

[0100]

[0101] Where V is the reserve information of materials to be processed, V i1 is the i1th material reserve semantic vector in the material reserve information to be processed, CXL is the material reserve feature word vector set, CXL i1 and CXL i1 are the i1th and j1th material reserve feature word vectors in the material reserve feature word vector set, N1 is the number of material reserve feature word vectors in the material reserve feature word vector set, δ1 is the first weight parameter, ∈ is the material reserve semantic deviation vector information, ∈ i1 It is the i1th material reserve semantic deviation vector in the material reserve semantic deviation vector information.

[0102] It should be noted that the first weight parameter and the material reserve semantic deviation vector information may be set by the user or obtained based on historical data, and the embodiments of the present invention do not limit this.

[0103] It should be noted that through the material reserve semantic computing model, the semantics of the entire material reserve information can be comprehensively considered to construct an overall semantic representation of each material reserve, which helps to obtain a more accurate semantic representation of material reserves, and can effectively improve the correlation analysis of material reserve information and the overall performance of the model, so that subsequent analysis (such as material reserve correlation analysis, knowledge graph construction, etc.) can better understand and process the semantic content of material reserve data.

[0104] It should be noted that the first weight parameter ranges from [0.1 to 2] and determines the importance of material reserve information in its calculation. A larger first weight parameter results in a greater influence on the weighted similarity of the semantic vector calculation results, thereby giving greater contribution to material reserve terms with higher semantic similarity. Assigning a larger first weight parameter to material reserve information with higher similarity, such as between [1.1, 2], highlights the relationship between semantically similar materials and is more suitable for scenarios with clear relationships between material reserves. When the first weight parameter takes a smaller value, such as between [0.1, 1.1], the model more evenly distributes the contributions of each feature word, reducing the influence of feature words with higher similarity, and is more suitable for scenarios with more diverse and complex material reserve information.

[0105] It should be noted that the material reserve semantic deviation vector information plays an adjustment and correction role in the above calculations. It can compensate for deviations in model training or fine-tune certain material reserve data to help the model adapt to the distribution of actual data. This can prevent the model from overfitting certain specific feature words, help the model better adapt to new data, and improve the stability and robustness of the model. For example, the following Table 1 shows an example of material reserve semantic deviation vector information:

[0106] Table 1 Example information of semantic deviation vector information of material reserves

[0107] Material characteristic words Material reserve semantic deviation vector (three-dimensional vector) Material A [0.1,0.4,0.3] Supplies B [0.3,0.8,0.2] Supplies C [0.2,0.1,0.5]

[0108] It can be seen that the implementation of the knowledge graph construction method based on material reserve information described in the embodiment of the present invention is conducive to realizing the correlation analysis of materials, thereby providing strong support for the rational allocation of materials and emergency management, and further improving the optimization efficiency of material reserves and supply chain management and the emergency response efficiency.

[0109] In an optional embodiment, the material reserve information to be processed and the material reserve training data set are processed to obtain material reserve sequence information, including:

[0110] S3221: Using the material reserve training data set, perform training processing on the first material reserve initial model information to obtain first material reserve training result information and first material reserve training model information;

[0111] It should be noted that the initial model information of the first material reserve is the BERT model.

[0112] The above processing is to use the material reserve training data set as the input of the first material reserve initial model information, train through the first material reserve initial model information, and obtain the first material reserve training result information, wherein the first material reserve training model information is a model obtained after training with the material reserve training data set.

[0113] S3222: Calculate and process the first material reserve training result information to obtain a first material reserve loss value;

[0114] S3223, determining whether the first material reserve loss value is less than a preset first material reserve loss threshold, and obtaining a first determination result;

[0115] It should be noted that the preset first material reserve loss threshold has a value range of [0.05, 0.2], which is not specifically limited in the embodiment of the present invention.

[0116] When the first judgment result is no, determining that the first material reserve training model information is the first material reserve initial model information, and reacquiring the material reserve training data set, executing S3221;

[0117] What needs to be processed is that the above-mentioned re-acquiring the material reserve training data set is to obtain a second material reserve training data set and determine the second material reserve training data set as the material reserve training data set. By updating the material reserve training data set, a more accurate first material reserve optimization model can be obtained.

[0118] When the first judgment result is yes, determining the first material reserve training model information as a first material reserve optimization model;

[0119] S3224, using the first material reserve optimization model, calculates and processes the material reserve information to be processed to obtain material reserve sequence information;

[0120] It should be noted that the calculation process described above uses the pending material reserve information as input to the first material reserve optimization model. This calculation, performed using the first material reserve optimization model, yields material reserve sequence information. This sequence is a set of material reserve sequences arranged by time, status, or other relevant dimensions. This sequence reflects the state changes of material reserves at different times or under different conditions and is often used for subsequent analysis, forecasting, optimization, and decision-making.

[0121] It can be seen that the implementation of the knowledge graph construction method based on material reserve information described in the embodiment of the present invention is conducive to realizing the correlation analysis of materials, thereby providing strong support for the rational allocation of materials and emergency management, and further improving the optimization efficiency of material reserves and supply chain management and the emergency response efficiency.

[0122] In an optional embodiment, calculating and processing the first material reserve training result information to obtain the first material reserve loss value includes:

[0123] Using the first material reserve loss function calculation model, the first material reserve training result information is calculated and processed to obtain the first material reserve loss value;

[0124] Among them, the calculation model of the first material reserve loss function is:

[0125]

[0126] Where WZSS is the loss value of the first material reserve, YY is the first material reserve true label information corresponding to the first material reserve training result information, YYZ is the first material reserve training result information, M2 and N2 are the number of category information in the first material reserve training result information and the number of first material reserve training result values ​​corresponding to the category information in the first material reserve training result information, respectively. j i and YYZ j i They are respectively the i2th first material reserve true label value of the j2th category in the first material reserve true label information corresponding to the category information in the first material reserve training result information and the i2th first material reserve training result value of the j2th category in the category information in the first material reserve training result information, and δ2 is the second weight parameter.

[0127] It should be noted that the second weight parameter may be set by a user or obtained based on historical data, and the embodiment of the present invention does not limit this.

[0128] It should be noted that the first material reserve loss function calculation model adjusts model parameters to enable the material reserve model to more accurately predict material reserve information based on a given training dataset. This can help improve inventory management, supply chain configuration, demand forecasting, and other aspects.

[0129] It should be noted that the value range of the second weight parameter is [0.1, 3], which is used to control the contribution of each part in the calculation model of the first material reserve loss function. Adjusting this parameter helps optimize the performance of the model and ensures that the loss function not only focuses on prediction accuracy but also takes into account other factors such as category balance.

[0130] It can be seen that the implementation of the knowledge graph construction method based on material reserve information described in the embodiment of the present invention is conducive to realizing the correlation analysis of materials, thereby providing strong support for the rational allocation of materials and emergency management, and further improving the optimization efficiency of material reserves and supply chain management and the emergency response efficiency.

[0131] In an optional embodiment, the material reserve sequence information is processed to obtain material reserve relationship information, including:

[0132] S3231, using a material reserve similarity calculation model, calculating and processing the material reserve sequence information to obtain material reserve similarity information;

[0133] Among them, the material reserve similarity calculation model is:

[0134]

[0135] Where, XSD is the material reserve similarity information, XSD i3,j3 is the material reserve similarity value between the i3th material and the j3th material in the material reserve similarity information, WZCB i3 and WZCB j3 are the i3th and j3th material reserve sequences in the material reserve sequence information, |·| represents the modulus of the vector, and θ1 represents the first similarity offset parameter;

[0136] It should be noted that the first similarity offset parameter may be set by a user or obtained based on historical data, and this is not limited in the embodiment of the present invention.

[0137] It should be noted that the material reserve similarity calculation model can measure the similarity between two material reserve sequences, thereby helping to understand the relationship between different material reserves.

[0138] It should be noted that the value range of the first similarity offset parameter is between [1, 5], which can make the obtained material reserve similarity information more in line with business needs. For example, in some cases it is necessary to strengthen the similarity of certain features, while in other cases it may be necessary to weaken them.

[0139] S3232, performing category matching processing on the material reserve similarity information to obtain material reserve category information;

[0140] It should be noted that the above-mentioned category matching process can be performed using algorithms or tools such as K-means, hierarchical clustering or DBSCAN, and the specific embodiments of the present invention do not limit this.

[0141] It should be noted that through category matching processing, material reserves can be managed more effectively, ensuring that required materials can be dispatched quickly and accurately when needed, thereby improving response efficiency.

[0142] S3233: Calculate and process the material reserve category information to obtain material reserve relationship information.

[0143] It can be seen that the implementation of the knowledge graph construction method based on material reserve information described in the embodiment of the present invention is conducive to realizing the correlation analysis of materials, thereby providing strong support for the rational allocation of materials and emergency management, and further improving the optimization efficiency of material reserves and supply chain management and the emergency response efficiency.

[0144] In an optional embodiment, the material reserve category information is calculated and processed to obtain material reserve relationship information, including:

[0145] S32331, obtain material reserve category dataset;

[0146] It's important to note that the dataset used to train and validate machine learning models aims to learn the categorical characteristics of material reserve information, enabling prediction, classification, or optimal decision-making for new data. This dataset contains not only basic information about the materials but also the corresponding classification labels for each material, allowing the model to learn how to assign material categories based on different characteristics.

[0147] S32332: Using the material reserve category data set, training the initial model information of the material reserve category to obtain material reserve category training result information and material reserve category training model information;

[0148] It should be noted that the initial model information of the material reserve category is the support vector machine.

[0149] It should be noted that the above-mentioned training process uses the material reserve category data set as the input of the material reserve category initial model information, and trains through the material reserve category initial model information to obtain the material reserve category training result information, wherein the material reserve category training model information is the model obtained after training with the material reserve category data set.

[0150] S32333, calculating and processing the material reserve category training result information to obtain the material reserve category loss value;

[0151] S32334, determining whether the material reserve category loss value is less than a preset material reserve category loss threshold, and obtaining a second determination result;

[0152] It should be noted that the preset material reserve category loss threshold value ranges from [0.01, 0.05], and is not specifically limited in the embodiment of the present invention.

[0153] It should be noted that the preset first material reserve loss threshold value ranges from [0.05, 0.2], and the preset material reserve category loss threshold value ranges from [0.01, 0.05]. This adjustment strategy from "rough" to "fine" can effectively balance the exploration and convergence in the training process, so that as the technical solution of this application deepens, while maintaining sufficient exploration capabilities, it can converge to a better solution in the later stage, avoiding premature convergence or overfitting.

[0154] When the second judgment result is no, it is determined that the material reserve category training model information is the material reserve category initial model information, and S32331 is executed;

[0155] When the second judgment result is yes, determining the material reserve category training model information as the material reserve category optimization model information;

[0156] S32335, using the material reserve category optimization model information, calculates and processes the material reserve category information to obtain material reserve relationship information.

[0157] It should be noted that the calculation process described above uses material reserve category information as input to the material reserve category optimization model. This calculation generates material reserve relationship information. This material reserve relationship information includes information about the connections, dependencies, optimal configuration plans, and resource allocation strategies between materials. This relationship information plays a crucial role in material reserve management, particularly in emergency management, material allocation, and supply chain management.

[0158] It can be seen that the implementation of the knowledge graph construction method based on material reserve information described in the embodiment of the present invention is conducive to realizing the correlation analysis of materials, thereby providing strong support for the rational allocation of materials and emergency management, and further improving the optimization efficiency of material reserves and supply chain management and the emergency response efficiency.

[0159] In an optional embodiment, the material reserve category training result information is calculated and processed to obtain the material reserve category loss value, including:

[0160] Using the material reserve category calculation model, the material reserve category training result information is calculated and processed to obtain the material reserve category loss value;

[0161] Among them, the calculation model for material reserve categories is:

[0162]

[0163] Where LBSS is the loss value of material reserve category, LBZ(x i4 ) is the training result value of the i4th sample in the material reserve category training result information, LB(x i4 ) is the true result value of the preset i4th sample, N4 is the number of training result values ​​in the material reserve category training result information, and LBPC is the preset material reserve category deviation coefficient.

[0164] It should be noted that the preset material reserve category deviation coefficient may be set by the user or obtained based on historical data, and the embodiment of the present invention does not limit this.

[0165] It's important to note that the material reserve classification calculation model guides the model optimization process by calculating and weighting the error between the predicted results and the true labels. Setting the coefficient of deviation and squared error helps adjust and optimize the material reserve model, thereby improving its accuracy and stability in the material reserve field.

[0166] It should be noted that the preset deviation coefficient of the material reserve category ranges from [0.1, 0.3]. It can slightly adjust the deviation of the model, allowing the model to maintain flexibility within a larger range and adapt to different data distributions while avoiding overfitting. This affects whether the model can effectively correct the prediction error and make the model output closer to the true label.

[0167] It can be seen that the implementation of the knowledge graph construction method based on material reserve information described in the embodiment of the present invention is conducive to realizing the correlation analysis of materials, thereby providing strong support for the rational allocation of materials and emergency management, and further improving the optimization efficiency of material reserves and supply chain management and the emergency response efficiency.

[0168] In an optional embodiment, the material reserve relationship information and the material classification information are processed to obtain material reserve knowledge graph information, including:

[0169] S331, processing the material reserve relationship information and the material classification information to obtain material classification sequence information;

[0170] It should be noted that the above processing can be performed using algorithms or tools such as K-means and decision trees, and the specific embodiments of the present invention do not limit this.

[0171] It should be noted that through the above processing, meaningful material classification sequence information can be extracted from large-scale material reserve data, thereby improving the efficiency and accuracy of material classification and management.

[0172] S332, fusing the material classification sequence information and the material reserve relationship information to obtain a material classification matrix;

[0173] S333: Integrate and process the material classification matrix to obtain material reserve knowledge graph information.

[0174] It should be noted that the above integration processing can be performed through DeepWalk, Node2Vec or GraphSAGE, and the specific embodiment of the present invention does not limit it.

[0175] It should be noted that through integration and processing, a material reserve knowledge graph can be constructed. These graphs provide information on materials, material categories, and material relationships, thereby providing strong support for the rational allocation of materials and emergency management, thereby improving the optimization efficiency of material reserves and supply chain management and the efficiency of emergency response.

[0176] It can be seen that the knowledge graph construction method based on the material reserve information described in the embodiment of the present application is beneficial to realize the correlation analysis of the materials, thereby providing strong support for the reasonable allocation and emergency management of the materials, and further improving the optimization efficiency and emergency response efficiency of the material reserve and supply chain management.

[0177] In an optional embodiment, the material classification sequence information and the material reserve relationship information are fused to obtain a material classification matrix, including:

[0178] S3321, the material classification sequence information is vectorized to obtain material classification vector information;

[0179] It should be noted that the above vectorization processing can be processed using Word2Vec, TF-IDF and other algorithms or tools. Specifically, the embodiment of the present application is not limited.

[0180] It should be noted that through the vectorization processing, the similarity and semantic relationship between the material classifications can be effectively captured, providing accurate classification and clustering capabilities, and ultimately optimizing the material management and supply chain efficiency.

[0181] S3322, using a material classification matrix calculation model, the material classification vector information and the material reserve relationship information are calculated to obtain a material classification matrix;

[0182] The material classification matrix calculation model is:

[0183]

[0184] In the formula, WJZ is the material classification matrix, K is the material reserve relationship information, L is the material classification vector information, K i5,j5 is the relationship vector of the i5th material and the j5th material in the material reserve relationship information, L j5 is the classification vector of the j5th material in the material classification vector information, N5 is the number of relationship vectors in the material reserve relationship information, and M5 is the number of classification vectors in the material classification vector information.

[0185] It should be noted that through the material classification matrix calculation model, the relationship between the material reserve and the classification can be understood, which provides support for optimizing the material management, supply chain scheduling and emergency response. Through weighted summation and logarithmic transformation, it can be avoided to excessively bias certain specific material categories, and the balance and stability of the results can be ensured.

[0186] It can be seen that the knowledge graph construction method based on the material reserve information described in the embodiment of the present application is beneficial to realize the correlation analysis of the materials, thereby providing strong support for the reasonable allocation and emergency management of the materials, and further improving the optimization efficiency and emergency response efficiency of the material reserve and supply chain management.

[0187] Example 2

[0188] See also Figure 2 , Figure 2 This is a schematic diagram of a knowledge graph construction device based on material reserve information disclosed in an embodiment of the present invention. Figure 2 The described knowledge graph construction device based on material reserve information is applied to a knowledge graph construction optimization system based on material reserve information, such as a local server or cloud server for constructing a knowledge graph based on material reserve information, and the embodiment of the present invention does not limit this. Figure 2 As shown, the knowledge graph construction device based on material reserve information includes:

[0189] Acquisition module 201, used to acquire material reserve information and material classification information;

[0190] A preprocessing module 202 is used to preprocess the material reserve information to obtain preprocessed material reserve information;

[0191] The calculation module 203 is used to process the pre-processed material reserve information and material classification information to obtain material reserve knowledge graph information.

[0192] It can be seen that the implementation of the knowledge graph construction device based on material reserve information described in the embodiment of the present invention is conducive to realizing the correlation analysis of materials, thereby providing strong support for the reasonable allocation of materials and emergency management, and further improving the optimization efficiency of material reserves and supply chain management and the emergency response efficiency.

[0193] Example 3

[0194] See also Figure 3 , Figure 3 This is a schematic diagram of another structure of a knowledge graph construction device based on material reserve information disclosed in an embodiment of the present invention. Figure 3 The described knowledge graph construction device based on material reserve information is applied to a knowledge graph construction optimization system based on material reserve information, such as a local server or cloud server for constructing a knowledge graph based on material reserve information, and the embodiment of the present invention does not limit this. Figure 3 As shown, the knowledge graph construction device based on material reserve information includes:

[0195] Processor 301;

[0196] A memory 302 coupled to the processor 301 and storing executable program code;

[0197] The processor 301 calls the executable program code stored in the memory 302 to execute part or all of the steps of the method for constructing a knowledge graph based on material reserve information in Example 1.

[0198] It can be seen that the implementation of the knowledge graph construction device based on material reserve information described in the embodiment of the present invention is conducive to realizing the correlation analysis of materials, thereby providing strong support for the reasonable allocation of materials and emergency management, and further improving the optimization efficiency of material reserves and supply chain management and the emergency response efficiency.

[0199] Example 4

[0200] An embodiment of the present invention discloses a computer-readable storage medium, which stores computer instructions. When the computer instructions are called, they are used to execute some or all steps of the method for constructing a knowledge graph based on material reserve information in embodiment one.

[0201] Example 5

[0202] An embodiment of the present invention discloses a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to enable a computer to execute some or all of the steps in the method for constructing a knowledge graph based on material reserve information described in Example 1.

[0203] The system embodiments described above are merely illustrative. Modules described as separate components may or may not be physically separate, and components shown as modules may or may not be physical modules. They may be located in one place or distributed across multiple network modules. Some or all of these modules may be selected based on actual needs to achieve the objectives of this embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0204] Through the detailed description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus the necessary general hardware platform, or of course, by means of hardware. Based on this understanding, the above technical solution, in essence, or the portion that contributes to the prior art, can be embodied in the form of a software product, which can be stored in a computer-readable storage medium, including a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

[0205] Finally, it should be noted that the method and device for constructing a knowledge graph based on material reserve information disclosed in the embodiment of the present invention only disclose a preferred embodiment of the present invention, which is only used to illustrate the technical solution of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, ordinary technicians in this field should understand that it is still possible to modify the technical solutions recorded in the aforementioned embodiments, or to make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for constructing a knowledge graph based on material reserve information, characterized in that: The method comprises: S1, obtain material reserve information and material classification information; S2, preprocessing the material reserve information to obtain preprocessed material reserve information; S3: Process the pre-processed material reserve information and the material classification information to obtain material reserve knowledge graph information.

2. The method for constructing a knowledge graph based on material reserve information according to claim 1, characterized in that: The preprocessing of the material reserve information to obtain preprocessed material reserve information includes: S21, removing stop words from the material reserve information to obtain first material reserve preprocessing information; S22, removing special characters from the first material reserve preprocessing information to obtain second material reserve preprocessing information; S23, performing formatting uniform processing on the second material reserve preprocessing information to obtain third material reserve preprocessing information; S24, performing synonym conversion processing on the third material reserve preprocessing information to obtain preprocessed material reserve information.

3. The method for constructing a knowledge graph based on material reserve information according to claim 1, characterized in that: The processing of the pre-processed material reserve information and the material classification information to obtain material reserve knowledge graph information includes: S31, obtaining a material reserve training data set; S32, processing the material reserve training data set and the pre-processed material reserve information to obtain material reserve relationship information; S33: Process the material reserve relationship information and the material classification information to obtain material reserve knowledge graph information.

4. The method for constructing a knowledge graph based on material reserve information according to claim 3, characterized in that: The processing of the material reserve training data set and the pre-processed material reserve information to obtain material reserve relationship information includes: S321, processing the pre-processed material reserve information to obtain unprocessed material reserve information; S322, processing the to-be-processed material reserve information and the material reserve training data set to obtain material reserve sequence information; S323: Process the material reserve sequence information to obtain material reserve relationship information.

5. The method for constructing a knowledge graph based on material reserve information according to claim 4, characterized in that: The processing of the pre-processed material reserve information to obtain the material reserve information to be processed includes: S3211, performing feature word extraction processing on the pre-processed material reserve information to obtain material reserve feature word information; S3212: Process the material reserve feature word information and the pre-processed material reserve information to obtain a material reserve feature word vector set; S3213, using a material reserve semantic calculation model, calculating and processing the material reserve feature word vector set to obtain material reserve information to be processed; The material reserve semantic computing model is: Where, V is the reserve information of materials to be processed, V i1 is the i1th material reserve semantic vector in the material reserve information to be processed, CXL is the material reserve feature word vector set, CXL i1 and CXL j1 are the i1th and j1th material reserve feature word vectors in the material reserve feature word vector set, N1 is the number of material reserve feature word vectors in the material reserve feature word vector set, δ1 is the first weight parameter, ∈ is the material reserve semantic deviation vector information, ∈ i1 It is the i1th material reserve semantic deviation vector in the material reserve semantic deviation vector information.

6. The method for constructing a knowledge graph based on material reserve information according to claim 4, characterized in that: The material reserve sequence information is processed to obtain material reserve relationship information, including: S3231, using a material reserve similarity calculation model, calculating and processing the material reserve sequence information to obtain material reserve similarity information; The material reserve similarity calculation model is: Where, XSD is the similarity information of the material reserves, XSD i3,j3 is the material reserve similarity value between the i3th material and the j3th material in the material reserve similarity information, WZCB i3 and WZCB j3 are the i3th and j3th material reserve sequences in the material reserve sequence information, respectively, |·| represents the modulus of the vector, and θ1 represents the first similarity offset parameter; S3232, performing category matching processing on the material reserve similarity information to obtain material reserve category information; S3233: Calculate and process the material reserve category information to obtain material reserve relationship information.

7. The method for constructing a knowledge graph based on material reserve information according to claim 3, characterized in that: The material reserve relationship information and the material classification information are processed to obtain material reserve knowledge graph information, including: S331, processing the material reserve relationship information and the material classification information to obtain material classification sequence information; S332, fusing the material classification sequence information and the material reserve relationship information to obtain a material classification matrix; S333: Integrate the material classification matrix to obtain material reserve knowledge graph information.

8. A knowledge graph construction device based on material reserve information, characterized in that: The device comprises: Acquisition module, used to obtain material reserve information and material classification information; A preprocessing module, configured to preprocess the material reserve information to obtain preprocessed material reserve information; The calculation module is used to process the pre-processed material reserve information and the material classification information to obtain material reserve knowledge graph information.

9. A knowledge graph construction device based on material reserve information, characterized in that: The device comprises: processor; a memory coupled to the processor and storing executable program code; The processor calls the executable program code stored in the memory to execute the knowledge graph construction method based on material reserve information as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, which, when called, are used to execute the knowledge graph construction method based on material reserve information as described in any one of claims 1 to 7.

Images