Resource allocation method and system based on AI artificial intelligence matrix

Abstract

The invention belongs to the technical field of artificial intelligence, and particularly relates to a resource allocation method and system based on an AI artificial intelligence matrix, and the method comprises the steps: 1, scanning a resource file, and determining a file type; 2, an optimal allocation scheme is adapted based on the extracted features; step 3, performing data classified storage based on the optimal allocation scheme; 4, extracting a resource file; and step 5, AI autonomous learning and conversion. The resource allocation system is used for implementing the resource allocation method and comprises a first unit, a second unit, a third unit, a fourth unit and a fifth unit. The new resource allocation method and system can be constructed based on the AI, resource allocation can be improved from underlying logic, files are stored according to data types, storage architecture is simplified, storage pressure is reduced, meanwhile, the AI can autonomously learn according to file data, once the files are lost, the files can be autonomously restored according to file fragments, and the file allocation efficiency is improved. The loss is reduced.

Description

Technical Field

[0001] The present invention belongs to the field of artificial intelligence technology, and specifically relates to a resource allocation method and system based on an AI artificial intelligence matrix. Background Art

[0002] Artificial intelligence (AI) is a technology that simulates human intelligence, enabling computers to perform tasks that normally require human intelligence. Artificial intelligence is a technology that enables computers and machines to simulate human intelligence and problem-solving abilities. It can be used independently or in combination with other technologies to perform tasks that would otherwise require human intelligence or human intervention. Machine learning and deep learning are the core technologies of AI, which use neural networks to learn from data and make predictions or classifications. These algorithms can automatically extract features and achieve unsupervised learning. AI has a wide range of applications, covering speech recognition, computer vision, autonomous driving and other fields. It is changing the way people live and work, improving efficiency and quality.

[0003] At present, artificial intelligence is widely used in various fields, and is also involved in the field of resource allocation. However, since the allocation methods and systems currently installed on computers are based on the allocation logic before the emergence of AI, even with the intervention of AI, it is difficult to fundamentally affect resource allocation. Therefore, although the impact of AI intervention on resource allocation exists, it is small and can only assist allocation to a certain extent, with limited efficiency improvement. Moreover, at this stage, once the source file of a resource is damaged or lost during the storage process, it is difficult to retrieve it, which poses certain security risks and is not conducive to the long-term storage and retrieval of resource files. Summary of the invention

[0004] The purpose of the present invention is to provide a resource allocation method and system based on an AI artificial intelligence matrix, which can build a new resource allocation method and system based on AI, improve resource allocation from the underlying logic, store files separately according to data types, simplify the storage architecture, and reduce storage pressure. At the same time, AI can learn autonomously based on file data. Once a file is lost, it can be restored autonomously based on file fragments to reduce losses.

[0005] The technical solution adopted by the present invention is as follows: A resource allocation method based on an AI artificial intelligence matrix, comprising: Scan resource files to determine file types. During the resource file storage and allocation process, extract the features of each type of content in the file, and annotate these features with type labels and component labels based on the file. Through feature representation, convert the type labels and component labels into quantifiable data, and apply principal component analysis to simplify the data results. Adapting the optimal allocation plan based on the extracted features, using a linear regression or classification algorithm to predict the optimal allocation plan of resources, wherein the linear regression or classification algorithm can automatically adjust the resource allocation strategy according to historical data and current status; Classify and store data based on the optimal allocation scheme, use the optimal allocation scheme to classify and process the various types of content in the file, optimize the storage architecture, perform multi-level storage in parallel, and perform capability labeling, convert the capability labeling into quantifiable data again, and simplify the data by applying principal component analysis; Resource file extraction: during the resource file extraction process, based on the component tags, the various types of file data of the resource file are re-extracted, and then reintegrated and outputted; AI autonomously learns and transforms. During the process of resource file storage, allocation, and extraction, AI continuously deepens its learning based on the file type and content data, extracts advanced features, and its neural network and back-propagation algorithm automatically generate nonlinear modeling and build a multi-layer neural network. It uses the back-propagation algorithm for parameter optimization, which can improve the ability to accurately generate similar complete content based on fragmented file content, and assist in restoring damaged file content when part of the file is lost.

[0006] As a preferred solution, the resource file is scanned to determine the file type, and the file is identified according to the size of the file and at least one of the characteristic data of the file to obtain an identification result, wherein the identification result includes whether the file is one or more of a picture file, a text file, a table file, a formula file, and a video file; The picture files, text files, table files, formula files and video files are annotated with corresponding type tags and component tags. The type tags can be used to identify the file type, and the component tags can be used to identify which parent file different types of sub-files belong to.

[0007] As a preferred solution, the principal component analysis is used to identify and simplify patterns in the data, generate new coordinate axes by linearly combining the original variables, and sort these coordinate axes according to their importance; The mathematical form of the principal component analysis describes the calculation steps as follows: Step 1. Set the original variable , …、 The n-times observation data matrix is: ; Step 2: Normalize the data matrix by column, and the normalized data matrix is ​​still recorded as X; Step 3. Find the correlation coefficient matrix is defined as: ; Step 4. Find the characteristic root λ1≥λ2≥λp>0 of the characteristic equation of R det(R−λE)=0; Step 5: Determine the number of main components , Determined according to actual problems, generally 70-90% is extracted; Step 6. Calculate m corresponding one-way eigenvectors: ; Step 7. Calculate the principal components: .

[0008] As a preferred solution, the optimal allocation solution is adapted based on the extracted features, and different storage algorithms are matched based on the type labels of the image files, text files, table files, formula files, and video files in the extracted features to form the optimal allocation solution; In forming the optimal allocation plan, the file format is compared and analyzed with the existing algorithms, the algorithm is scored according to the corresponding algorithm and the corresponding file type, and the score of the optimal allocation plan is predicted. Finally, the allocation plan with the highest output score is the optimal allocation plan.

[0009] As a preferred solution, the scoring gradient descent formula of the optimal allocation solution is: During the calculation process, iterate continuously until the termination condition is reached; Among them, p represents the position or parameter, q represents the query or another parameter, α is the learning rate, which is used to control the update speed, β is the momentum coefficient, which is used to accelerate the convergence process, and e represents the base of the natural exponential function, which is often used to represent the exponential function. is a constant.

[0010] As a preferred solution, the resource file extraction creates a container based on the target storage and starts the container; Mounting the preset resource collection program, resource collection component label and type label directory in the container; Running the resource collection program, searching the container for target resource files of the component label and the type label through the resource collection program, wherein the target resource files include any one or more of the following: picture files, text files, table files, formula files, and video files; The target resource file is copied to the resource collection directory and restored and integrated using AI to output a complete target resource file.

[0011] As a preferred solution, the AI ​​autonomous learning and transformation includes: A model learning module is used to unsupervisedly learn the behavior patterns of the input, output and storage algorithms of the resource files, and gradually form a resource generation data model with correlations among the three, obtain model parameters of built-in related data, and calibrate with the resource generation data model to gradually improve it; The real-time data acquisition module generates the data model based on the resource obtained by unsupervised learning, collates the data generated by the resource generation data model according to the subsequent input and output resource file data, and self-iterates and updates the resource file data based on the real-time input and output resource file data.

[0012] As a preferred solution, the AI ​​autonomous learning and transformation also includes: The external learning module absorbs relevant resource data models based on the Internet of Things, and performs data cleaning, feature extraction, and model training on the relevant data models, and continuously forms self-iteration updates through online learning or incremental learning.

[0013] As a preferred solution, the AI ​​autonomous learning and transformation also includes: The resource storage capacity monitoring and autonomous allocation module can monitor the remaining storage capacity of the storage in real time, and optimize and divide the remaining storage capacity according to the remaining storage capacity and the proportion of storage space occupied by historical storage of image files, text files, table files, formula files and video files, so as to adapt to different storage needs in real time.

[0014] A resource allocation system based on an AI artificial intelligence matrix, used to implement the above-mentioned resource allocation method, comprising: Unit 1, in the process of resource file storage and allocation, extracts the features of each type of content in the file, and labels the features with type labels and component labels based on the file, converts the type labels and component labels into quantifiable data through feature representation, and simplifies the data results by applying principal component analysis; The second unit uses a linear regression or classification algorithm to predict the optimal allocation plan of resources, and the linear regression or classification algorithm can automatically adjust the resource allocation strategy according to historical data and current status; The third unit uses the optimal allocation scheme to classify and process the various types of content in the file, optimize the storage architecture, perform multi-level storage in parallel, and perform capability labeling, and convert the capability labeling into quantifiable data again, and simplify the data by applying principal component analysis; The fourth unit, in the process of extracting the resource file, re-extracts the file data of each type of the resource file based on the component tags, and then re-integrates and outputs the data; Unit 5: During the resource file storage allocation and extraction process, AI continuously deepens its learning based on the file type and content data, extracts advanced features, and its neural network and back propagation algorithm automatically generate nonlinear modeling and build a multi-layer neural network. It uses the back propagation algorithm for parameter optimization, which can improve the ability to accurately generate similar complete content based on the fragmented content of the file, and assist in restoring the damaged file content when part of the file is lost.

[0015] The technical effects achieved by the present invention are: The present invention constructs a new resource allocation method and system based on AI, which can improve resource allocation from the bottom-level logic, store the traditional resource file content by type, and perform algorithm matching according to the file type, so that different resource files can match the optimal allocation plan, improve allocation efficiency, and simplify the storage architecture and reduce storage pressure by storing files separately according to data types. Moreover, when a certain type of file is damaged, other file types will not be affected, so that when a file is lost, the degree of damage is reduced and the resource files are preserved more completely.

[0016] The present invention allows AI to learn autonomously based on file data, absorbs relevant resource data models based on the Internet of Things, and performs data cleaning, feature extraction, and model training on the relevant data models. It continuously forms self-iteration updates through online learning or incremental learning. When files are lost, they can be restored autonomously based on file fragments, and the restored content can be 80%-100% close to the original content. Simulation replaces the original data to assist in data restoration, thereby reducing damage and achieving the purpose of restoring the true content of the data. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a flow chart of a resource allocation method based on an AI artificial intelligence matrix of the present invention; Figure 2 It is a schematic diagram of the process of extracting resource files in the resource allocation method based on the AI ​​artificial intelligence matrix of the present invention; Figure 3 It is a structural diagram of AI autonomous learning and transformation in the resource allocation method based on AI artificial intelligence matrix of the present invention; Figure 4 It is a structural diagram of the resource allocation system based on the AI ​​artificial intelligence matrix of the present invention. DETAILED DESCRIPTION

[0018] In order to make the purpose and advantages of the present invention more clearly understood, the present invention is specifically described below in conjunction with embodiments. It should be understood that the following text is only used to describe one or several specific embodiments of the present invention, and does not strictly limit the scope of protection of the specific claims of the present invention.

[0019] like Figure 1-Figure 3 As shown, a resource allocation method based on an AI artificial intelligence matrix includes: Step 1: Scan resource files to determine the file type; Step 2: Adapt the optimal allocation scheme based on the extracted features; Step 3: Classify and store data based on the optimal allocation plan; Step 4: Extract resource files; Step 5: AI autonomous learning and transformation; The above five steps can maximize the efficiency of resource allocation, simplify the storage architecture, and reduce storage pressure. At the same time, AI can learn autonomously based on file data. Once a file is lost, it can restore it autonomously based on file fragments to reduce losses.

[0020] Step 1: Scan resource files to determine file types. During the resource file storage and allocation process, extract the features of each type of content in the file, and annotate these features with type labels and component labels based on the file. Through feature representation, convert type labels and component labels into quantifiable data, and use principal component analysis to simplify the data results.

[0021] The resource file scanning to determine the file type is to identify the file according to at least one of the file size and the file feature data to obtain an identification result, wherein the identification result includes whether the file is one or more of a picture file, a text file, a table file, a formula file, and a video file; Image files, text files, table files, formula files and video files are marked with corresponding type labels and component labels. Type labels can be used to identify file types, and component labels can be used to identify which parent file different types of sub-files belong to. The significance of type labels is to classify and store file contents for easy identification, while component labels are used to re-extract and combine classified and stored files.

[0022] In order to simplify the data structure and improve computational efficiency, principal component analysis (PCA) is used to identify and simplify patterns in the data, generate new coordinate axes by linearly combining the original variables, and sort these axes by their importance; The mathematical form of principal component analysis describes the calculation steps as follows: Step 1. Set the original variable , …、 The n-times observation data matrix is: ; Step 2: Normalize the data matrix by column, and the normalized data matrix is ​​still recorded as X; Step 3. Find the correlation coefficient matrix is defined as: ; Step 4. Find the characteristic root λ1≥λ2≥λp>0 of the characteristic equation of R det(R−λE)=0; Step 5: Determine the number of main components , Determined according to actual problems, generally 70-90% is extracted, and 90% is extracted in this embodiment; Step 6. Calculate m corresponding one-way eigenvectors: ; Step 7. Calculate the principal components: .

[0023] Principal component analysis (PCA) maps the original high-dimensional data to a low-dimensional space through orthogonal transformation, significantly reducing the dimension of the data, thereby simplifying the subsequent data processing and model training process. This can not only improve computational efficiency, but also reduce the risk of overfitting, making the model more robust, thereby making the data feature extraction and storage model of type labels and component labels more stable, improving extraction and labeling efficiency, while principal component analysis (PCA) can remove noise and redundant information in the data and improve the signal-to-noise ratio of the data. This is of great significance for improving the accuracy and reliability of data analysis. By removing unimportant variables and variables with strong correlation, principal component analysis (PCA) can help better focus on key information.

[0024] Step 2: Adapt the optimal allocation plan based on the extracted features and use linear regression or classification algorithms to predict the optimal allocation plan for resources. Linear regression or classification algorithms can automatically adjust resource allocation strategies based on historical data and current status.

[0025] Among them, the optimal allocation scheme based on the extracted features is to match different storage algorithms based on the type labels of the image files, text files, table files, formula files, and video files in the extracted features to form the optimal allocation scheme; In forming the optimal allocation plan, the file format is compared and analyzed with the existing algorithms, the algorithm is scored according to the corresponding algorithm and the corresponding file type, and the score of the optimal allocation plan is predicted. Finally, the allocation plan with the highest output score is the optimal allocation plan.

[0026] Among them, the storage algorithms of image files, text files, table files, formula files, and video files are all existing algorithms in the prior art, for example: Image file storage algorithms include: BMP (bitmap), GIF (Graphics Interchange Format), JPEG / JPG, PNG (Portable Network Graphics), SVG (Scalable Vector Graphics), TIFF (Tagged Image File Format), etc.

[0027] Text file storage algorithms include: ASCII encoding, UTF-8 encoding, Huffman encoding, Base64 encoding and various compression algorithms.

[0028] Table file storage algorithms include: CSV (Comma Separated Value) format, TSV (Tab Separated Value) format, XLSX (Excel XML) format, ODS (OpenDocument Spreadsheet) format, SQL database, HDF5 format, Parquet format, ORC (Optimized Row Columnar) format, etc.

[0029] Formula file storage algorithms include: TeX (LaTeX) format, MathML format, OMML format, ContentMathML format, AsciiMath format, OpenDocument Formula format, etc.

[0030] Video file storage algorithms include: H.264 (MPEG-4 AVC) encoding, H.265 (HEVC) encoding, VP9 encoding, AV1 encoding, MPEG-2 encoding, Theora encoding, DivX encoding, Sorenson Spark encoding, RealVideo encoding, Indeo encoding, etc.

[0031] In practical applications, how to choose a suitable algorithm requires algorithm recommendation through matrix decomposition, for example: The matrix describes the scores (1-5 points) of 5 network evaluations (U1, U2, U3, U4, U5) on 4 algorithms (D1, D2, D30, D4). - means no score. The current goal is to predict those without scores, and then recommend the optimal algorithm for the corresponding file based on the predicted scores.

[0032] The matrix table is as follows: D1 D2 D3 D4 U1 5 3 - 1 U2 4 - - 1 U3 1 1 - 5 U4 1 - - 4 U5 - 1 5 4 The matrix R can be approximately expressed as the product of P and Q: .

[0033] So the evaluation matrix can be calculated as follows: .

[0034] Here’s how to do it: Step 1. Calculate the value of each element of the reconstructed matrix and list the expression: .

[0035] Step 2: Construct a loss function, using the square of the difference between the original matrix element and the new matrix element as the loss function. Therefore, the square expression of the difference between each segment of elements is: ; The final loss function is the sum of all known item losses: .

[0036] Step 3, use gradient descent to obtain p and q components: According to the gradient descent formula: ; Update the variable in the direction of the negative gradient: ; Continue iterating until the termination condition is reached.

[0037] Of course, in order to prevent overfitting, a regularization term needs to be added. Just change the loss value to the following formula: ; At the same time, the gradient descent formula is updated to: During the calculation process, iterate continuously until the termination condition is reached; Among them, p represents the position or parameter, q represents the query or another parameter, α is the learning rate, which is used to control the update speed, β is the momentum coefficient, which is used to accelerate the convergence process, and e represents the base of the natural exponential function, which is often used to represent the exponential function. is a constant.

[0038] Step 3: Classify and store data based on the optimal allocation plan. Use the optimal allocation plan to classify and store each type of content in the file, optimize the storage architecture, perform multi-level storage in parallel, and perform capability labeling. Convert the capability labels into quantifiable data again, and simplify the data by reapplying the principal component analysis in step 1.

[0039] Step 4: Resource file extraction. During the resource file extraction process, based on the component tags, various types of file data of the resource file are re-extracted, and then reintegrated and output.

[0040] It mainly includes the following steps: Step 1, resource file extraction creates a container based on the target storage and starts the container; Step 2. Mount the preset resource collection program, resource collection component label, and type label directory in the container; Step 3, run the resource collection program, and search the target resource files of the component label and the type label in the container through the resource collection program. The target resource files include any one or more of the following: picture files, text files, table files, formula files and video files; Step 4. Copy the target resource file to the resource collection directory and use AI to restore and integrate it to output the complete target resource file.

[0041] Step 5: AI autonomous learning and transformation. During the process of resource file storage, allocation and extraction, AI continuously deepens its learning based on the file type and content data, extracts advanced features, and its neural network and back propagation algorithm automatically generate nonlinear modeling and build a multi-layer neural network. The back propagation algorithm is used for parameter optimization to improve the ability to accurately generate similar complete content based on the fragmented content of the file, so as to assist in restoring the damaged file content when part of the file is lost.

[0042] Among them, AI autonomous learning and transformation include: Model learning module, used for unsupervised learning of resource file input, output and storage algorithm behavior patterns, and gradually forming a resource generation data model with correlation among the three, obtaining model parameters with built-in relevant data, and checking with the resource generation data model to gradually improve it; The real-time data acquisition module generates a data model based on the resources obtained through unsupervised learning, collates the data generated by the resource generation data model with the subsequent input and output resource file data, and iterates itself based on the real-time input and output resource file data.

[0043] In order to continuously optimize the AI ​​model, AI autonomous learning and transformation also includes: The external learning module absorbs relevant resource data models based on the Internet of Things, and performs data cleaning, feature extraction, and model training on the relevant data models, and continuously forms self-iteration updates through online learning or incremental learning.

[0044] To optimize the storage architecture, AI autonomous learning and transformation also include: The resource storage capacity monitoring and autonomous allocation module can monitor the remaining storage capacity of the storage in real time, and optimize and divide the remaining storage capacity according to the remaining storage capacity and the proportion of storage space occupied by historical storage of image files, text files, table files, formula files and video files, so as to adapt to different storage needs in real time.

[0045] like Figure 1-Figure 4 As shown, a resource allocation system based on an AI artificial intelligence matrix is ​​used to implement the above-mentioned resource allocation method, including: Unit 1: In the process of resource file storage and allocation, the features of each type of content in the file are extracted, and type labels and component labels are annotated based on the file. Through feature representation, the type labels and component labels are converted into quantifiable data, and principal component analysis is applied to simplify the data results. Unit 2: Use linear regression or classification algorithms to predict the optimal resource allocation plan. Linear regression or classification algorithms can automatically adjust resource allocation strategies based on historical data and current status. Unit 3: Use the optimal allocation scheme to classify and process the various types of content in the file, optimize the storage architecture, parallelize multi-level storage, and perform capability labeling. Then, the capability labels are converted into quantifiable data, and the data is simplified by applying principal component analysis. Unit 4, in the process of resource file extraction, re-extracts the various types of file data of the resource file based on the component tags, and then re-integrates and outputs them; Unit 5: During the resource file storage allocation and extraction process, AI continuously deepens its learning based on the file type and content data, extracts advanced features, and its neural network and back propagation algorithm automatically generate nonlinear modeling and build a multi-layer neural network. It uses the back propagation algorithm for parameter optimization, which can improve the ability to accurately generate similar complete content based on the fragmented content of the file, and assist in restoring the damaged file content when part of the file is lost.

[0046] The above is only a preferred embodiment of the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made without departing from the principles of the present invention, and these improvements and modifications should also be considered as the protection scope of the present invention. The structures, devices and operating methods not specifically described and explained in the present invention shall be implemented according to the conventional means in the art unless otherwise specified and limited.

Claims

1. A resource allocation method based on AI artificial intelligence matrix, characterized in that: include: Scan resource files to determine file types. During the resource file storage and allocation process, extract the features of each type of content in the file, and annotate these features with type labels and component labels based on the file. Through feature representation, convert the type labels and component labels into quantifiable data, and apply principal component analysis to simplify the data results. Adapting the optimal allocation plan based on the extracted features, using a linear regression or classification algorithm to predict the optimal allocation plan of resources, wherein the linear regression or classification algorithm can automatically adjust the resource allocation strategy according to historical data and current status; Classify and store data based on the optimal allocation scheme, use the optimal allocation scheme to classify and process the various types of content in the file, optimize the storage architecture, perform multi-level storage in parallel, and perform capability labeling, convert the capability labeling into quantifiable data again, and simplify the data by applying principal component analysis; Resource file extraction: during the resource file extraction process, based on the component tags, the various types of file data of the resource file are re-extracted, and then reintegrated and outputted; AI autonomously learns and transforms. During the process of resource file storage, allocation, and extraction, AI continuously deepens its learning based on the file type and content data, extracts advanced features, and its neural network and back-propagation algorithm automatically generate nonlinear modeling and build a multi-layer neural network. It uses the back-propagation algorithm for parameter optimization, which can improve the ability to accurately generate similar complete content based on fragmented file content, and assist in restoring damaged file content when part of the file is lost.

2. The resource allocation method based on AI artificial intelligence matrix according to claim 1, characterized in that: Scan the resource file to determine the file type, and identify the file according to the size of the file and at least one of the characteristic data of the file to obtain an identification result, wherein the identification result includes whether the file is one or more of a picture file, a text file, a table file, a formula file, and a video file; The picture files, text files, table files, formula files and video files are annotated with corresponding type tags and component tags. The type tags can be used to identify the file type, and the component tags can be used to identify which parent file different types of sub-files belong to.

3. The resource allocation method based on AI artificial intelligence matrix according to claim 1, characterized in that: The principal component analysis is used to identify and simplify patterns in the data, generate new coordinate axes by linearly combining the original variables, and rank these coordinate axes according to their importance; The mathematical form of the principal component analysis describes the calculation steps as follows: Step 1. Set the original variable , …、 The n-times observation data matrix is: ； Step 2: Normalize the data matrix by column, and the normalized data matrix is ​​still recorded as X; Step 3. Find the correlation coefficient matrix is defined as: ； Step 4. Find the characteristic root λ1≥λ2≥λp>0 of the characteristic equation of R det(R−λE)=0; Step 5: Determine the number of main components , Determined according to actual problems, generally 70-90% is extracted; Step 6. Calculate m corresponding one-way eigenvectors: ； Step 7. Calculate the principal components: 。 4. The resource allocation method based on AI artificial intelligence matrix according to claim 1, characterized in that: Adapt the optimal allocation plan based on the extracted features, and match different storage algorithms based on the type labels of image files, text files, table files, formula files, and video files in the extracted features to form the optimal allocation plan; In forming the optimal allocation plan, the file format is compared and analyzed with the existing algorithms, the algorithm is scored according to the corresponding algorithm and the corresponding file type, and the score of the optimal allocation plan is predicted. Finally, the allocation plan with the highest output score is the optimal allocation plan.

5. The resource allocation method based on AI artificial intelligence matrix according to claim 1, characterized in that: The scoring gradient descent formula of the optimal allocation scheme is: During the calculation process, iterate continuously until the termination condition is reached; Among them, p represents the position or parameter, q represents the query or another parameter, α is the learning rate, which is used to control the update speed, β is the momentum coefficient, which is used to accelerate the convergence process, and e represents the base of the natural exponential function, which is often used to represent the exponential function. is a constant.

6. The resource allocation method based on AI artificial intelligence matrix according to claim 1, characterized in that: The resource file extraction creates a container based on the target storage and starts the container; Mounting the preset resource collection program, resource collection component label and type label directory in the container; Running the resource collection program, searching the container for target resource files of the component label and the type label through the resource collection program, wherein the target resource files include any one or more of the following: picture files, text files, table files, formula files, and video files; The target resource file is copied to the resource collection directory and restored and integrated using AI to output a complete target resource file.

7. The resource allocation method based on AI artificial intelligence matrix according to claim 1 is characterized in that: The AI ​​autonomous learning and transformation includes: The model learning module is used to unsupervisedly learn the behavior patterns of the input, output and storage algorithms of the resource files, and gradually form a resource generation data model with correlations among the three, obtain model parameters of built-in related data, and calibrate with the resource generation data model to gradually improve it; The real-time data acquisition module generates the data model based on the resource obtained by unsupervised learning, collates the data generated by the resource generation data model according to the subsequent input and output resource file data, and self-iterates and updates the resource file data based on the real-time input and output resource file data.

8. The resource allocation method based on AI artificial intelligence matrix according to claim 1 is characterized in that: The AI ​​autonomous learning and transformation also includes: The external learning module absorbs relevant resource data models based on the Internet of Things, and performs data cleaning, feature extraction, and model training on the relevant data models, and continuously forms self-iteration updates through online learning or incremental learning.

9. The resource allocation method based on AI artificial intelligence matrix according to claim 1 is characterized in that: The AI ​​autonomous learning and transformation also includes: The resource storage capacity monitoring and autonomous allocation module can monitor the remaining storage capacity of the storage in real time, and optimize and divide the remaining storage capacity according to the remaining storage capacity and the proportion of storage space occupied by historical storage of image files, text files, table files, formula files and video files, so as to adapt to different storage needs in real time.

10. A resource allocation system based on AI artificial intelligence matrix, used to implement the resource allocation method according to any one of claims 1 to 9, characterized in that: include: Unit 1: In the process of resource file storage and allocation, the features of each type of content in the file are extracted, and type labels and component labels are annotated on these features based on the file. The type labels and component labels are converted into quantifiable data through feature representation, and principal component analysis is applied to simplify the data results. Unit 2, using linear regression or classification algorithms to predict optimal resource allocation solutions, wherein the linear regression or classification algorithms can automatically adjust resource allocation strategies based on historical data and current status; The third unit uses the optimal allocation scheme to classify and process the various types of content in the file, optimize the storage architecture, perform multi-level storage in parallel, and perform capability labeling, and convert the capability labeling into quantifiable data again, and simplify the data by applying principal component analysis; The fourth unit, in the process of resource file extraction, re-extracts the file data of each type of the resource file based on the component tags, and then re-integrates and outputs the data; Unit 5: During the resource file storage allocation and extraction process, AI continuously deepens its learning based on the file type and content data, extracts advanced features, and its neural network and back propagation algorithm automatically generate nonlinear modeling and build a multi-layer neural network. It uses the back propagation algorithm for parameter optimization, which can improve the ability to accurately generate similar complete content based on the fragmented content of the file, and assist in restoring the damaged file content when part of the file is lost.

Images