Metadata generation method and system for heterogeneous system

Abstract

The invention relates to the technical field of data processing, in particular to a metadata generation method and system for a heterogeneous system, and the method comprises the steps: obtaining a metadata sequence of the heterogeneous system and a corresponding metadata access sequence, calculating a metadata popularity sequence corresponding to the metadata sequence based on the metadata sequence and the metadata access sequence; and predicting a metadata popularity sequence at a next preset moment based on the metadata popularity sequence, and classifying the metadata popularity sequence at the next preset moment based on a preset popularity threshold to obtain a first metadata popularity sequence and a second metadata popularity sequence. According to the method, by predicting and classifying the heat of the metadata, the system can optimize the index of the metadata, so that the retrieval speed is increased, the metadata is classified and clustered according to the heat sequence, data which is frequently accessed can be stored in a centralized mode, waste of storage resources is reduced, and the storage layout is optimized.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method and system for generating metadata for a heterogeneous system. Background Art

[0002] Heterogeneous systems refer to systems composed of different types of hardware, software, or networks. These systems are usually used to process diverse data sources and computing tasks to improve system performance and flexibility. These systems may have different architectures, platforms, data formats, and communication protocols.

[0003] Traditional methods do not have an indexing mechanism optimized for popularity, resulting in long data retrieval time; and the data is not classified and clustered in traditional methods, resulting in non-optimal allocation of storage resources; and traditional methods cannot dynamically adjust the index and storage layout according to data access patterns and popularity changes; and traditional methods have relatively fixed management and organization of data structures and cannot effectively cope with frequently changing data environments. Summary of the invention

[0004] The purpose of the present invention is to address the problems existing in the background technology and to propose a method and system for generating metadata for heterogeneous systems.

[0005] The technical solution of the present invention is a method for generating metadata for a heterogeneous system, comprising:

[0006] Acquire a metadata sequence and a corresponding metadata access sequence of a heterogeneous system, and calculate a metadata heat sequence corresponding to the metadata sequence based on the metadata sequence and the metadata access sequence;

[0007] Predicting a metadata heat sequence at a next preset moment based on the metadata heat sequence, and classifying the metadata heat sequence at the next preset moment based on a preset heat threshold to obtain a first metadata heat sequence and a second metadata heat sequence;

[0008] Acquire a first metadata sequence and a second metadata sequence based on the first metadata heat sequence and the second metadata heat sequence, and cluster the first metadata sequence and the second metadata sequence to obtain a plurality of first clustering clusters corresponding to the first metadata sequence and a plurality of second clustering clusters corresponding to the second metadata sequence;

[0009] Performing feature extraction on the plurality of first clusters and the plurality of second clusters to obtain a plurality of first data features and a plurality of second data features, and calculating a first similarity between the plurality of first data features and a second similarity between the plurality of second data features by using a similarity model;

[0010] sorting the plurality of first clusters according to the first similarity to obtain a first cluster sequence, sorting the plurality of second clusters according to the second similarity to obtain a second cluster sequence, and combining the first cluster sequence and the second cluster sequence to obtain a cluster sequence;

[0011] Based on the order of the cluster sequence, a B+ tree is used to establish indexes of each cluster with file marks as keywords.

[0012] Preferably, the heat calculation formula of the metadata is as follows:

[0013]

[0014] Among them, H j represents the popularity of the metadata when the jth access request is received, m represents the access time of the mth most recent metadata access when the jth access request is received, and t j represents the access time when the metadata is accessed for the jth time, t j-m Indicates the access time when jm accessed the metadata for the first time.

[0015] Preferably, predicting the metadata heat sequence at the next preset moment based on the metadata heat sequence includes:

[0016] The metadata heat sequence is accumulated to obtain an accumulated metadata heat sequence, and the accumulated metadata heat sequence expression is as follows:

[0017] h (1 )=[h (1) (k1),h (1) (k2),...,h (1 )(k i ),...,h( 1) (k n )];

[0018] Among them, h (1) Represents the cumulative metadata heat sequence, h( 1 )(k i ) represents the ith accumulated metadata heat in the accumulated metadata heat sequence, and h( 0 )(k j ) represents the jth metadata heat in the metadata heat sequence, Δk j represents the spacing of metadata heat sequences, and Δk j =k j -k j-1 ;

[0019] A grayscale prediction model is constructed for the accumulated metadata heat sequence, and the grayscale prediction model is as follows:

[0020]

[0021] Among them, a represents the development coefficient and b represents the gray action.

[0022] Preferably, predicting the metadata heat sequence at the next preset moment based on the metadata heat sequence further includes:

[0023] A whitening response formula is constructed based on the grayscale prediction model, and the whitening response formula is as follows:

[0024]

[0025] in, represents the whitening response of the grayscale prediction model when t = k1, and and Indicates that the accumulated metadata heat sequence is in the interval [k i ,k i+1 ] background value, and and

[0026] The metadata heat sequence is restored based on the grayscale prediction model to obtain the metadata heat sequence at the next preset moment. The metadata heat sequence at the next preset moment is expressed as follows:

[0027]

[0028] in, Indicates the metadata heat sequence at the next preset time.

[0029] Preferably, clustering the first metadata sequence and the second metadata sequence to obtain a plurality of first clusters corresponding to the first metadata sequence and a plurality of second clusters corresponding to the second metadata sequence includes:

[0030] A state value function is predefined, and the clustering stop point of the first metadata sequence and the second metadata sequence is determined by the state value function. The value function expression is as follows:

[0031]

[0032] Among them, δ represents the state value function, θ represents, η represents the Euclidean distance, and χ p represents a data object including a data attribute p, represents the data center of the association rule u, where u represents a certain association rule, P represents the data attribute set, and p represents the data attribute;

[0033] Calculating the attribution degree of each element of the first metadata sequence and the second metadata sequence based on a fuzzy mean algorithm and a relationship matrix;

[0034] An objective function is constructed with the minimum degree of belonging as the goal, constraints are set for the state value function, and the objective function and constraints are solved by the Lagrange multiplier method to obtain a plurality of first clusters corresponding to the first metadata sequence and a plurality of second clusters corresponding to the second metadata sequence, wherein the constraints are as follows:

[0035]

[0036] Among them, st represents the constraint condition, h represents the data to be clustered, and u ph It represents the correlation between the data to be clustered h and the data attribute p, and ε represents the total number of data to be clustered.

[0037] Preferably, feature extraction is performed on the plurality of first clusters and the plurality of second clusters to obtain a plurality of first data features and a plurality of second data features, and a first similarity between the plurality of first data features and a second similarity between the plurality of second data features is calculated by a similarity model, including:

[0038] For each of the first clusters and each of the first clusters, extracting a first feature vector of the first cluster and a second feature vector of the second cluster by a principal component analysis method;

[0039] A first similarity between each pair of the first feature vectors and a second similarity between each pair of the second feature vectors are calculated using a similarity model to obtain a first similarity sequence and a second similarity sequence.

[0040] Preferably, sorting a plurality of first clusters according to the first similarity to obtain a first cluster sequence, and sorting a plurality of second clusters according to the second similarity to obtain a second cluster sequence, and combining the first cluster sequence and the second cluster sequence to obtain a cluster sequence, including:

[0041] Sorting the first similarity sequence and the second similarity sequence to obtain a first similarity order sequence and a second similarity order sequence;

[0042] sorting a plurality of first clusters based on the first similarity order sequence to obtain a first cluster sequence, and sorting a plurality of second clusters based on the second similarity order sequence to obtain a second cluster sequence;

[0043] The second clustering sequence is combined into the first clustering sequence, and the first element in the second clustering sequence is combined with the last element in the first clustering sequence to obtain a clustering sequence.

[0044] Preferably, the similarity model adopts cosine similarity.

[0045] Preferably, based on the order of the cluster sequence, a B+ tree is used to establish indexes of each cluster using file marks as keywords, including:

[0046] Determine the order of the B+ tree, and initialize the root node of the B+ tree to be empty;

[0047] Inserting the file mark of each cluster as a keyword into the B+ tree, and in the process of inserting, updating the inner node, if the inner node exceeds the maximum capacity, splitting, and ascending the intermediate keyword to the parent node, and creating a new inner node;

[0048] The file mark of the cluster is stored in the leaf node of the B+ tree, and the leaf node is kept sorted by keyword. If the leaf node exceeds the maximum capacity, the leaf node is split, and the intermediate keyword is raised to the parent node, and a new leaf node is created;

[0049] After the cluster sequence traversal is completed, a new B+ tree is obtained, and indexes of each cluster are established based on the new B+ tree.

[0050] The technical solution of the present invention is a system for generating metadata for heterogeneous systems, which is applicable to the method for generating metadata for heterogeneous systems, comprising:

[0051] A data acquisition unit, the data acquisition unit is used to acquire a metadata sequence and a corresponding metadata access sequence of a heterogeneous system, and calculate a metadata heat sequence corresponding to the metadata sequence based on the metadata sequence and the metadata access sequence;

[0052] a heat prediction unit, the heat prediction unit being used to predict a metadata heat sequence at a next preset moment based on the metadata heat sequence, and classify the metadata heat sequence at the next preset moment based on a preset heat threshold to obtain a first metadata heat sequence and a second metadata heat sequence;

[0053] a data clustering unit, the data clustering unit being used to obtain a first metadata sequence and a second metadata sequence based on the first metadata heat sequence and the second metadata heat sequence, and cluster the first metadata sequence and the second metadata sequence to obtain a plurality of first clustering clusters corresponding to the first metadata sequence and a plurality of second clustering clusters corresponding to the second metadata sequence;

[0054] a similarity calculation unit, the similarity calculation unit being used to perform feature extraction on the plurality of first clusters and the plurality of second clusters to obtain a plurality of first data features and a plurality of second data features, and to calculate a first similarity between the plurality of first data features and a second similarity between the plurality of second data features through a similarity model;

[0055] A cluster combination unit, the cluster combination unit is used to sort the multiple first clusters according to the size of the first similarity to obtain a first cluster sequence, and sort the multiple second clusters according to the size of the second similarity to obtain a second cluster sequence, and combine the first cluster sequence and the second cluster sequence to obtain a cluster sequence;

[0056] An index building unit is used to build indexes of each cluster respectively based on the order of the cluster sequence by using a B+ tree and taking a file mark as a keyword.

[0057] Compared with the prior art, the above technical solution of the present invention has the following beneficial technical effects:

[0058] The present invention predicts and classifies metadata heat, so that the system can optimize the indexing of metadata, thereby accelerating the retrieval speed, and classifies and clusters metadata according to heat sequences, which helps to centrally store frequently accessed data, reduce the waste of storage resources, and optimize the storage layout. Through classification and sorting based on heat sequences, it can respond to users' needs for frequently accessed data more quickly, and use the B+ tree structure to establish an index, which can effectively support dynamic updates and efficient data access. Through clustering metadata and extracting features, the internal structure and relationship of the data can be revealed, which helps to more effectively organize and manage metadata and improve data availability and consistency. The order of clustering clusters and the index structure can be dynamically adjusted according to changes in data, thereby being able to adapt to different access modes and data changes. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 A schematic flow chart of the steps of an overall method in an embodiment of the present invention;

[0060] Figure 2 A schematic diagram of the system architecture of an overall system in an embodiment of the present invention.

[0061] Figure numerals: 1. Data acquisition unit; 2. Heat prediction unit; 3. Data clustering unit; 4. Similarity calculation unit; 5. Cluster combination unit; 6. Index establishment unit. DETAILED DESCRIPTION

[0062] Embodiment 1, as Figure 1 As shown, the present invention proposes a method for generating metadata for a heterogeneous system, comprising:

[0063] S1. Obtain the metadata sequence and the corresponding metadata access sequence of the heterogeneous system, and calculate the metadata heat sequence corresponding to the metadata sequence based on the metadata sequence and the metadata access sequence;

[0064] S2, predicting the metadata heat sequence at the next preset moment based on the metadata heat sequence, and classifying the metadata heat sequence at the next preset moment based on a preset heat threshold to obtain a first metadata heat sequence and a second metadata heat sequence;

[0065] S3, acquiring a first metadata sequence and a second metadata sequence based on the first metadata heat sequence and the second metadata heat sequence, and clustering the first metadata sequence and the second metadata sequence to obtain a plurality of first clustering clusters corresponding to the first metadata sequence and a plurality of second clustering clusters corresponding to the second metadata sequence;

[0066] S4, performing feature extraction on the plurality of first clusters and the plurality of second clusters to obtain a plurality of first data features and a plurality of second data features, and calculating a first similarity between the plurality of first data features and a second similarity between the plurality of second data features through a similarity model;

[0067] S5, sorting the plurality of first clusters according to the first similarity to obtain a first cluster sequence, sorting the plurality of second clusters according to the second similarity to obtain a second cluster sequence, and combining the first cluster sequence and the second cluster sequence to obtain a cluster sequence;

[0068] S6. Based on the order of the cluster sequence, a B+ tree is used with the file mark as a keyword to establish the index of each cluster.

[0069] In the present invention, metadata sequence refers to an information sequence that describes the structure, attributes or relationships of data; metadata access sequence refers to a sequence that records access records to metadata; metadata heat sequence refers to a metadata importance sequence calculated based on access frequency; heat threshold refers to a preset boundary value used to classify metadata heat sequence; clustering refers to the process of grouping data into several similar data sets; B+ tree refers to a balanced tree data structure used for efficient data retrieval and sorting; index refers to a structure that organizes data to accelerate queries.

[0070] Embodiment 2, a method for generating metadata for a heterogeneous system proposed by the present invention, compared with embodiment 1, this embodiment further includes: a heat calculation formula for metadata is as follows:

[0071]

[0072] Among them, H j represents the popularity of the metadata when the jth access request is received, m represents the access time of the mth most recent metadata access when the jth access request is received, and t j represents the access time when the metadata is accessed for the jth time, t j-m Indicates the access time when jm accessed the metadata for the first time.

[0073] In an optional embodiment, predicting the metadata heat sequence at the next preset moment based on the metadata heat sequence includes:

[0074] A1. Accumulate the metadata heat sequence to obtain the accumulated metadata heat sequence. The accumulated metadata heat sequence expression is as follows:

[0075] h (1) =[h (1) (k1),h (1) (k2),...,h (1) (k i ),...,h (1) (k n )];

[0076] Among them, h (1) Indicates the cumulative metadata heat sequence, h (1) (k i ) represents the ith accumulated metadata heat in the accumulated metadata heat sequence, and h (0) (k j ) represents the jth metadata heat in the metadata heat sequence, Δk j represents the spacing of metadata heat sequences, and Δk j =k j -k j-1 ;

[0077] A2. Construct a grayscale prediction model for the accumulated metadata heat sequence. The grayscale prediction model is as follows:

[0078]

[0079] Among them, a represents the development coefficient and b represents the gray action.

[0080] It should be noted that the grayscale prediction model refers to a prediction method based on gray system theory, which is used to process incomplete information and make predictions by modeling and analyzing the changing trends of data; the development coefficient means that in the grayscale prediction model, the development coefficient reflects the trend of data growth and helps determine the predicted value of future data; the gray action quantity refers to the quantity that measures the cumulative effect of data changes, and is used in the grayscale model to analyze and calculate data trends and fluctuations.

[0081] In an optional embodiment, predicting the metadata heat sequence at the next preset moment based on the metadata heat sequence further includes:

[0082] A3. Construct a whitening response formula based on the grayscale prediction model. The whitening response formula is as follows:

[0083]

[0084] in, represents the whitening response of the grayscale prediction model when t = k1, and and Indicates that the accumulated metadata heat sequence is in the interval [k i ,k i+1 ] background value, and and

[0085] A4. The metadata heat sequence is restored based on the grayscale prediction model to obtain the metadata heat sequence at the next preset moment. The metadata heat sequence at the next preset moment is expressed as follows:

[0086]

[0087] in, Indicates the metadata heat sequence at the next preset time.

[0088] It should be noted that whitening aims to convert data into white noise. In this case, the covariance matrix of the data becomes a unit matrix. The purpose of whitening is to remove the correlation in the data so that the data has more uniform statistical properties. Responsiveness usually refers to the response of data to certain inputs or external factors. In the whitening process, the concept of responsiveness may involve how to make the data processing or modeling process more sensitive or adaptable to different input changes.

[0089] In an optional embodiment, clustering the first metadata sequence and the second metadata sequence to obtain a plurality of first clusters corresponding to the first metadata sequence and a plurality of second clusters corresponding to the second metadata sequence includes:

[0090] B1. Predefine a state value function, and determine the clustering stop point of the first metadata sequence and the second metadata sequence through the state value function. The value function expression is as follows:

[0091]

[0092] Among them, δ represents the state value function, θ represents, η represents the Euclidean distance, and χ p represents a data object including a data attribute p, represents the data center of the association rule u, where u represents a certain association rule, P represents the data attribute set, and p represents the data attribute;

[0093] B2. Calculating the degree of belonging of each element of the first metadata sequence and the second metadata sequence based on a fuzzy mean algorithm and a relationship matrix;

[0094] B3. Construct an objective function with the minimum degree of belonging as the goal, set constraints on the state value function, and solve the objective function and the constraints by Lagrange multiplier method to obtain multiple first clusters corresponding to the first metadata sequence and multiple second clusters corresponding to the second metadata sequence. The constraints are as follows:

[0095]

[0096] Among them, st represents the constraint condition, h represents the data to be clustered, and u ph It represents the correlation between the data to be clustered h and the data attribute p, and ε represents the total number of data to be clustered.

[0097] In an optional embodiment, feature extraction is performed on the plurality of first clusters and the plurality of second clusters to obtain a plurality of first data features and a plurality of second data features, and a first similarity between the plurality of first data features and a second similarity between the plurality of second data features is calculated by a similarity model, including:

[0098] C1. For each first cluster and each second cluster, extract the first eigenvector of the first cluster and the second eigenvector of the second cluster by principal component analysis method;

[0099] C2. Calculate the first similarity between each pair of first feature vectors and the second similarity between each pair of second feature vectors through a similarity model to obtain a first similarity sequence and a second similarity sequence.

[0100] In an optional embodiment, sorting a plurality of first clusters according to the first similarity to obtain a first cluster sequence, sorting a plurality of second clusters according to the second similarity to obtain a second cluster sequence, and combining the first cluster sequence and the second cluster sequence to obtain a cluster sequence, including:

[0101] D1, sorting the first similarity sequence and the second similarity sequence to obtain a first similarity order sequence and a second similarity order sequence;

[0102] D2. Sort the plurality of first clusters based on the first similarity order sequence to obtain a first cluster sequence, and sort the plurality of second clusters based on the second similarity order sequence to obtain a second cluster sequence;

[0103] D3. Combining the second cluster sequence into the first cluster sequence, and combining the first element in the second cluster sequence with the last element in the first cluster sequence to obtain a cluster sequence.

[0104] In an optional embodiment, the similarity model uses cosine similarity.

[0105] It should be noted that cosine similarity is a measurement method for measuring the angle between two vectors in vector space. It is mainly used to evaluate the directional similarity of two vectors rather than their size or length. Cosine similarity is often used in text analysis and recommendation systems, especially when dealing with high-dimensional sparse data.

[0106] In an optional embodiment, based on the order of the cluster sequence, a B+ tree is used to establish indexes for each cluster using the file mark as a keyword, including:

[0107] E1. Determine the order of the B+ tree and initialize the root node of the B+ tree to be empty;

[0108] E2. Insert the file flag of each cluster as a keyword into the B+ tree, and during the insertion process, update the inner node. If the inner node exceeds the maximum capacity, split it, and move the intermediate keyword up to the parent node, and create a new inner node.

[0109] E3. Store the file flag of the cluster in the leaf node of the B+ tree and keep the leaf node sorted by keyword. If the leaf node exceeds the maximum capacity, split the leaf node, move the intermediate keyword up to the parent node, and create a new leaf node.

[0110] E4. After the cluster sequence traversal is completed, a new B+ tree is obtained, and the index of each cluster is established based on the new B+ tree.

[0111] It should be noted that the order is an important parameter in the B+ tree, which indicates how many child nodes (for internal nodes) or keywords (for leaf nodes) a node can contain at most; the root node is the top node of the B+ tree, the starting node of the tree, and when initialized, the root node is empty. When a keyword is inserted, a new node is created in the root node; insertion refers to adding a new keyword to the B+ tree; splitting refers to the operation of dividing a node into two nodes. When the internal node is split, part of the keywords are moved to the newly created internal node, and the middle keywords are promoted to the parent node. When the leaf node is split, the keywords are also divided into the new leaf node and kept sorted by keywords; during the splitting process of the B+ tree, the middle keywords are promoted to the parent node, and the parent node can continue to maintain the balance of the tree and provide a suitable split point for the child node; index is a data structure used to improve the efficiency of data retrieval. In the B+ tree, an index usually refers to a structure organized based on keywords (such as file flags) to help quickly find and access data.

[0112] Embodiment three, as Figure 2 As shown, the present invention proposes a metadata generation system for heterogeneous systems, which is applicable to the metadata generation method for heterogeneous systems, including:

[0113] Data acquisition unit 1, the data acquisition unit 1 is used to acquire the metadata sequence and the corresponding metadata access sequence of the heterogeneous system, and calculate the metadata heat sequence corresponding to the metadata sequence based on the metadata sequence and the metadata access sequence;

[0114] A heat prediction unit 2, the heat prediction unit 2 is used to predict the metadata heat sequence at the next preset moment based on the metadata heat sequence, and classify the metadata heat sequence at the next preset moment based on a preset heat threshold to obtain a first metadata heat sequence and a second metadata heat sequence;

[0115] The data clustering unit 3 is used to obtain a first metadata sequence and a second metadata sequence based on the first metadata heat sequence and the second metadata heat sequence, and cluster the first metadata sequence and the second metadata sequence to obtain a plurality of first clustering clusters corresponding to the first metadata sequence and a plurality of second clustering clusters corresponding to the second metadata sequence;

[0116] A similarity calculation unit 4 is used to extract features from the plurality of first clusters and the plurality of second clusters to obtain a plurality of first data features and a plurality of second data features, and calculate a first similarity between the plurality of first data features and a second similarity between the plurality of second data features through a similarity model;

[0117] The cluster combination unit 5 is used to sort the multiple first clusters according to the size of the first similarity to obtain a first cluster sequence, and sort the multiple second clusters according to the size of the second similarity to obtain a second cluster sequence, and combine the first cluster sequence and the second cluster sequence to obtain a cluster sequence;

[0118] The index building unit 6 is used to build the index of each cluster respectively based on the order of the cluster sequence by using the B+ tree and the file mark as the keyword.

[0119] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited thereto, and various changes can be made within the knowledge scope of technicians in the relevant technical field without departing from the purpose of the present invention.

Claims

1. A method for generating metadata for a heterogeneous system, characterized in that: include: Acquire a metadata sequence and a corresponding metadata access sequence of a heterogeneous system, and calculate a metadata heat sequence corresponding to the metadata sequence based on the metadata sequence and the metadata access sequence; Predicting a metadata heat sequence at a next preset moment based on the metadata heat sequence, and classifying the metadata heat sequence at the next preset moment based on a preset heat threshold to obtain a first metadata heat sequence and a second metadata heat sequence; Acquire a first metadata sequence and a second metadata sequence based on the first metadata heat sequence and the second metadata heat sequence, and cluster the first metadata sequence and the second metadata sequence to obtain a plurality of first clustering clusters corresponding to the first metadata sequence and a plurality of second clustering clusters corresponding to the second metadata sequence; Performing feature extraction on the plurality of first clusters and the plurality of second clusters to obtain a plurality of first data features and a plurality of second data features, and calculating a first similarity between the plurality of first data features and a second similarity between the plurality of second data features by using a similarity model; sorting the plurality of first clusters according to the first similarity to obtain a first cluster sequence, sorting the plurality of second clusters according to the second similarity to obtain a second cluster sequence, and combining the first cluster sequence and the second cluster sequence to obtain a cluster sequence; Based on the order of the cluster sequence, a B+ tree is used to establish indexes of each cluster with file marks as keywords.

2. The method for generating metadata for heterogeneous systems according to claim 1, characterized in that: The heat calculation formula of the metadata is as follows: Among them, H j represents the popularity of the metadata when the jth access request is received, m represents the access time of the m most recent metadata accesses when the jth access request is received, and t j represents the access time when the metadata is accessed for the jth time, t j-m Indicates the access time when jm accessed the metadata for the first time.

3. The method for generating metadata for heterogeneous systems according to claim 2, characterized in that: Predicting the metadata heat sequence at the next preset moment based on the metadata heat sequence includes: The metadata heat sequence is accumulated to obtain an accumulated metadata heat sequence, and the accumulated metadata heat sequence expression is as follows: h (1) =[h (1) (k1),h (1) (k2),...,h (1) (k i ),...,h (1) (k n )]; Among them, h (1) Indicates the cumulative metadata heat sequence, h (1) (k i ) represents the ith accumulated metadata heat in the accumulated metadata heat sequence, and h (0) (k j ) represents the jth metadata heat in the metadata heat sequence, Δk j represents the spacing of metadata heat sequence, and Δk j =k j -k j-1 ; A grayscale prediction model is constructed for the accumulated metadata heat sequence, and the grayscale prediction model is as follows: Among them, a represents the development coefficient and b represents the gray action.

4. The method for generating metadata for heterogeneous systems according to claim 3, characterized in that: Predicting the metadata heat sequence at the next preset moment based on the metadata heat sequence also includes: A whitening response formula is constructed based on the grayscale prediction model, and the whitening response formula is as follows: in, represents the whitening response of the grayscale prediction model when t = k1, and and Indicates that the accumulated metadata heat sequence is in the interval [k i ,k i+1 ] background value, and and The metadata heat sequence is restored based on the grayscale prediction model to obtain the metadata heat sequence at the next preset moment. The metadata heat sequence at the next preset moment is expressed as follows: in, Indicates the metadata heat sequence at the next preset time.

5. The method for generating metadata for heterogeneous systems according to claim 1, characterized in that: Clustering the first metadata sequence and the second metadata sequence to obtain a plurality of first clustering clusters corresponding to the first metadata sequence and a plurality of second clustering clusters corresponding to the second metadata sequence includes: A state value function is predefined, and the clustering stop point of the first metadata sequence and the second metadata sequence is determined by the state value function. The value function expression is as follows: Among them, δ represents the state value function, θ represents, η represents the Euclidean distance, and χ p represents a data object including a data attribute p, represents the data center of the association rule u, where u represents a certain association rule, P represents the data attribute set, and p represents the data attribute; Calculating the attribution degree of each element of the first metadata sequence and the second metadata sequence based on a fuzzy mean algorithm and a relationship matrix; An objective function is constructed with the minimum degree of belonging as the goal, constraints are set for the state value function, and the objective function and constraints are solved by the Lagrange multiplier method to obtain a plurality of first clusters corresponding to the first metadata sequence and a plurality of second clusters corresponding to the second metadata sequence, wherein the constraints are as follows: Among them, st represents the constraint condition, h represents the data to be clustered, and u ph It represents the correlation between the data to be clustered h and the data attribute p, and ε represents the total number of data to be clustered.

6. The method for generating metadata for heterogeneous systems according to claim 1, characterized in that: Extracting features from the plurality of first clusters and the plurality of second clusters to obtain a plurality of first data features and a plurality of second data features, and calculating a first similarity between the plurality of first data features and a second similarity between the plurality of second data features by using a similarity model, comprising: For each of the first clusters and each of the first clusters, extracting a first feature vector of the first cluster and a second feature vector of the second cluster by a principal component analysis method; A first similarity between each pair of the first feature vectors and a second similarity between each pair of the second feature vectors are calculated using a similarity model to obtain a first similarity sequence and a second similarity sequence.

7. The method for generating metadata for heterogeneous systems according to claim 6, characterized in that: Sorting a plurality of first clusters according to the first similarity to obtain a first cluster sequence, and sorting a plurality of second clusters according to the second similarity to obtain a second cluster sequence, and combining the first cluster sequence and the second cluster sequence to obtain a cluster sequence, including: Sorting the first similarity sequence and the second similarity sequence to obtain a first similarity order sequence and a second similarity order sequence; sorting a plurality of first clusters based on the first similarity order sequence to obtain a first cluster sequence, and sorting a plurality of second clusters based on the second similarity order sequence to obtain a second cluster sequence; The second clustering sequence is combined into the first clustering sequence, and the first element in the second clustering sequence is combined with the last element in the first clustering sequence to obtain a clustering sequence.

8. The method for generating metadata for heterogeneous systems according to claim 6, characterized in that: The similarity model adopts cosine similarity.

9. The method for generating metadata for heterogeneous systems according to claim 1, characterized in that: Based on the order of the cluster sequence, a B+ tree is used to establish indexes of each cluster using file marks as keywords, including: Determine the order of the B+ tree, and initialize the root node of the B+ tree to be empty; Inserting the file mark of each cluster as a keyword into the B+ tree, and in the process of inserting, updating the inner node, if the inner node exceeds the maximum capacity, splitting, and ascending the intermediate keyword to the parent node, and creating a new inner node; The file mark of the cluster is stored in the leaf node of the B+ tree, and the leaf node is kept sorted by keyword. If the leaf node exceeds the maximum capacity, the leaf node is split, and the intermediate keyword is raised to the parent node, and a new leaf node is created; After the cluster sequence traversal is completed, a new B+ tree is obtained, and indexes of each cluster are established based on the new B+ tree.

10. A metadata generation system for heterogeneous systems, which is applicable to the metadata generation method for heterogeneous systems according to any one of claims 1 to 9, characterized in that: include: A data acquisition unit (1), the data acquisition unit (1) is used to acquire a metadata sequence and a corresponding metadata access sequence of a heterogeneous system, and calculate a metadata heat sequence corresponding to the metadata sequence based on the metadata sequence and the metadata access sequence; A heat prediction unit (2), the heat prediction unit (2) being used to predict a metadata heat sequence at a next preset moment based on the metadata heat sequence, and to classify the metadata heat sequence at the next preset moment based on a preset heat threshold, so as to obtain a first metadata heat sequence and a second metadata heat sequence; A data clustering unit (3), the data clustering unit (3) is used to obtain a first metadata sequence and a second metadata sequence based on the first metadata heat sequence and the second metadata heat sequence, and cluster the first metadata sequence and the second metadata sequence to obtain a plurality of first clustering clusters corresponding to the first metadata sequence and a plurality of second clustering clusters corresponding to the second metadata sequence; A similarity calculation unit (4), the similarity calculation unit (4) is used to extract features from the plurality of first clusters and the plurality of second clusters to obtain a plurality of first data features and a plurality of second data features, and calculate a first similarity between the plurality of first data features and a second similarity between the plurality of second data features using a similarity model; A cluster combination unit (5), the cluster combination unit (5) is used to sort the multiple first clusters according to the size of the first similarity to obtain a first cluster sequence, and sort the multiple second clusters according to the size of the second similarity to obtain a second cluster sequence, and combine the first cluster sequence and the second cluster sequence to obtain a cluster sequence; An index building unit (6) is used to build indexes of each cluster respectively based on the order of the cluster sequence by using a B+ tree and taking a file mark as a keyword.

Images