A method and apparatus for automatic classification of data based on identification

Abstract

The application discloses a kind of based on identification's data automatic classification grading method and device, the method includes: using the classification label based on prediction model to the data to be predicted is classified, obtains the classification attribute of data identification;According to classification attribute and pre-set classification grading rule, obtain the security level based on label;Using keyword search, the corresponding relationship of keyword and security level is established;Judge whether hit keyword;If yes, take the security level corresponding to keyword, subtract the corresponding preset value, obtain joint grading result;If no, take the security level in type rule according to classification result, as grading result.The application is classified by using artificial intelligence prediction model preferentially, then through self-defined keyword search, cooperate type grading rule and keyword grading rule, classification grading is carried out jointly, on the basis of improving intelligent data classification grading, adapt to more business scenarios, provide effective support for data protection.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data classification, in particular to a data automatic classification method and device based on identification. BACKGROUND

[0002] Data classification is one of the core tasks of organizing data governance and data security, and is a means of integrating information security into data value and ensuring effective protection. However, due to different industries and enterprises, the classification and grading standards for data are different. At present, data security companies provide customized services based on the characteristics of industries or enterprises, and have not yet formed a classification and grading method with high universality. SUMMARY

[0003] The present application provides a data automatic classification method based on identification, defines a classification and grading security identification, classifies and predicts through an artificial intelligence model, and then matches the keyword grading rules and classification and grading rules to jointly obtain the classification and grading results of the data. The main technical problems to be solved are as follows: the current data classification and grading method is mostly based on the requirements of enterprises to provide customized services, which requires high labor cost. The current classification and grading method is not flexible, and often loses its original effect due to changes in enterprise management rules.

[0004] To achieve the above purpose, the present application adopts a data automatic classification method based on identification, which comprises the following steps:

[0005] S1: classifying the data to be predicted by using a classification label-based prediction model to obtain the classification attribute of the data identification;

[0006] S2: obtaining the security level based on the label according to the classification attribute and the preset classification and grading rules;

[0007] S3: keyword retrieval is adopted to establish the corresponding relationship between the keyword and the security level;

[0008] S4: it is judged whether the keyword is hit; if yes, the security level corresponding to the keyword is taken, the corresponding preset value is subtracted, and the joint grading result is obtained; if not, the security level in the type rule is taken as the grading result according to the classification result.

[0009] Optionally, before the step S1, the method further comprises: obtaining a sample file, extracting the features of the sample file, sending the features to a classifier model for training to obtain the prediction model.

[0010] Optionally, after obtaining the sample file, the method further comprises: performing word segmentation, stop word removal, keyword extraction, word matrix conversion, article matrix extraction and matrix dimension reduction processing on the sample file.

[0011] Optionally, the classifier model comprises one of an SVM model, a decision tree model, a K nearest neighbor model or a softmax model.

[0012] Optionally, the keyword retrieval comprises project name retrieval, department name retrieval or sensitive word retrieval.

[0013] Optionally, the corresponding preset value is a preset value corresponding to the label-based security level.

[0014] In order to achieve the above-mentioned purpose, the application further proposes a data automatic classification and grading device based on identification, the device comprises:

[0015] A classification module is configured to classify the to-be-predicted data by using a classification label-based prediction model, and obtain a classification attribute of the data identification.

[0016] An obtaining module is configured to obtain a label-based security level according to the classification attribute and a preset classification and grading rule.

[0017] An establishing module is configured to establish a corresponding relationship between a keyword and a security level by using keyword retrieval.

[0018] A grading module is configured to determine whether the keyword is hit, if yes, take the security level corresponding to the keyword, subtract a corresponding preset value to obtain a joint grading result, and if not, take the security level in the type rule according to the classification result as the grading result.

[0019] Compared with the prior art, the application has the beneficial effects that:

[0020] (1) An AI model is adopted to provide intelligent means for data classification.

[0021] (2) A configurable grading strategy is used to adapt to more variable business scenarios.

[0022] (3) An efficient and flexible classification and grading method is used to provide effective support for a series of governance measures such as data security protection and security audit. BRIEF DESCRIPTION OF DRAWINGS

[0023] Figure 1 is a flowchart of a data automatic classification and grading method based on identification of the application;

[0024] Figure 2 is a principle diagram of a data automatic classification and grading method based on identification of the application;

[0025] Figure 3 is a diagram of data classification based on machine learning;

[0026] Figure 4It is a schematic view of the identification-based data automatic classification and grading device.

[0027] The implementation, functional features and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0028] It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0029] In order to make the purpose, technical scheme and advantages of the present application more clear, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0030] As shown in the figure. According to the identification-based data automatic classification and grading method proposed by the present application, the following steps are included: Figure 1

[0031] S1: using a classification label-based prediction model to classify the data to be predicted, obtaining the classification attribute of the data identification;

[0032] S2: obtaining the label-based security level according to the classification attribute and the preset classification and grading rules;

[0033] S3: using keyword retrieval to establish the corresponding relationship between the keyword and the security level;

[0034] S4: judging whether the keyword is hit; if yes, taking the security level corresponding to the keyword, subtracting the corresponding preset value to obtain the joint grading result; if no, taking the security level in the type rule according to the classification result as the grading result.

[0035] In this embodiment, the overall idea of the identification-based data automatic classification and grading is to use an artificial intelligence prediction model to preferentially classify data, and then use self-defined keyword retrieval to jointly classify and grade by combining the type grading rules and the keyword grading rules.

[0036] 1. For classification and grading security identification:

[0037] (1) Data fingerprint: data unique identification, generally using the file MD5 value of the whole data as the fingerprint extraction method

[0038] (2) Classification attribute: each data should have specific type division, which is different according to specific business rules, industry attributes, enterprise requirements, etc.

[0039] (3) Grading attribute: each data should have specific security level division, which is used to implement specific protection strategy. ​

[0040] (4) Project name: According to specific business rules, the configured keywords to identify data and to grade, which can be project name, department name, etc.

[0041] 2. For the overall business process:

[0042] As shown in Figure 2 , the data automatic classification and grading method based on identification mainly includes four parts: machine learning model training, data discovery, data preprocessing, and intelligent classification and grading.

[0043] Model training: mainly through feature transformation and machine learning algorithm, based on a large number of sample data, training of classification prediction model containing labels.

[0044] Data discovery: through the timing task scanning personal terminal, file server, database and other data carriers, identifying various structured and unstructured data.

[0045] Data processing: including format conversion, data cleaning, feature extraction, fingerprint extraction, keyword search and other steps.

[0046] Intelligent classification and grading: through intelligent classification prediction, project name grading rules and classification grading rules matching, completing classification and grading, forming data security identification.

[0047] 3. For the data classification method based on machine learning:

[0048] As shown in Figure 3 , first of all, the method of artificial intelligence machine learning is adopted, a large number of sample data are used for supervised learning training, and a prediction model based on classification label is obtained. Through the model, the data to be predicted can be intelligently classified.

[0049] The core process of classification task based on machine learning:

[0050] (1) Sample file processing;

[0051] (2) Convert the preprocessed content into a feature matrix;

[0052] (3) Selectively reduce the dimension of the feature matrix;

[0053] (4) Send the feature matrix into the selected classifier model to train the model (classification result: label 1, label 2, label 3, label 4);

[0054] (5) The text to be predicted is processed in the same way and sent into the model in the last step.

[0055] 4. Grading rules based on classification results:

[0056] Based on the classification prediction method in 3, the classification attribute of the data identifier is obtained, and is matched with the classification grading rules in the following table, so that the security level based on the label can be obtained.

[0057] Classification result Security level Label 1 a Label 2 b Label 3 c Label 4 d

[0058] Result example:

[0059] Data fingerprint Classification result Security level dfd2sfga1t3sdfs Label 1 a opq2sf2a12sdfs Label 2 b zjd9sfyga23sdf Label 3 c ond2sf2a1sdfs2 Label 4 d

[0060] 5. Grading rules based on keyword search:

[0061] In actual business scenarios, it is often necessary to determine the level according to certain keywords (project name, department name.

[0062] Full-text keyword search is adopted to establish the corresponding relationship between the project name and the level, and the project name is only used as an example. Department name, special sensitive words, etc. can also be configured, as shown in the following table:

[0063] Project name Security level Name 1 A Name 2 B Name 3 C Name 4 D

[0064] Result example:

[0065] Data fingerprint Project name Security level dfd2sfga1t3sdfs Name 1 A opq2sf2a12sdfs Name 2 B zjd9sfyga23sdf null null ond2sf2a1sdfs2 Name 4 D

[0066] 6. Joint grading rules:

[0067] Grading rules:

[0068] 1. If the project name is not hit, the security level in the type rule is directly taken as the grading result according to the classification result.

[0069] 2. If the project name is hit, the security level corresponding to the project name is taken, and then the final grading result is obtained according to the type result and the value corresponding to it.

[0070] The rules are shown in the following table:

[0071]

[0072]

[0073] Example:

[0074]

[0075] Reference Figure 4 , Figure 4 The structure block diagram of the data automatic classification and grading device based on the identifier of the present application is shown in the following figure.

[0076] As Figure 4 shown, the data automatic classification and grading device based on the identifier provided by the embodiment of the present application comprises:

[0077] The classification module 10 is configured to classify the to-be-predicted data by using a prediction model based on a classification label, and obtain a classification attribute of the data identification;

[0078] The obtaining module 20 is configured to obtain a label-based security level according to the classification attribute and a preset classification grading rule;

[0079] The establishing module 30 is configured to establish a corresponding relationship between the keyword and the security level by using keyword retrieval;

[0080] The grading module 40 is configured to determine whether the keyword is hit; if yes, the security level corresponding to the keyword is taken, a preset value corresponding to the keyword is subtracted, and a joint grading result is obtained; if not, the security level in the type rule according to the classification result is taken as the grading result.

[0081] Other embodiments or specific implementations of the data automatic classification and grading device based on identification can refer to the above-mentioned method embodiments, and details are not described herein.

[0082] It should be noted that in this document, the terms "comprising", "containing", or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such a process, method, article or system. Without more limitations, the element defined by the statement "comprising a" does not exclude the presence of another identical element in the process, method, article or system including the element.

[0083] The above-mentioned embodiment numbers of the present application are only for description, and do not represent the advantages and disadvantages of the embodiments.

[0084] Through the above description of the embodiments, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be realized by software and necessary general hardware platform, of course, they can also be realized by hardware, but in many cases, the former is a better embodiment. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes a plurality of instructions for making a terminal device (which can be a mobile phone, computer, server, or network device, etc.) execute the methods described in various embodiments of the present application.

[0085] The above merely describes the preferred embodiments of the present application, and is not intended to limit the patent scope of the present application, and any equivalent structure or equivalent process conversion, or direct or indirect application in other related technical fields, which are made by using the content of the present application specification and drawings, are also included in the patent protection scope of the present application.

Claims

1. A method for automatic classification of data based on identification, characterized in that, The method comprises the following steps: S1: classifying the to-be-predicted data by using a prediction model based on a classification label to obtain a classification attribute of data identification; S2: obtaining a label-based security level according to the classification attribute and a preset classification grading rule; S3: performing keyword retrieval to establish a corresponding relationship between a keyword and a security level; S4: determining whether the keyword is hit; if yes, taking the security level corresponding to the keyword, subtracting a corresponding preset value corresponding to the label-based security level to obtain a joint grading result; if no, taking the security level in a type rule according to the classification result as a grading result.

2. The identification-based data auto-classification method of claim 1, wherein, Before step S1, the method further comprises: acquiring a sample file, extracting features of the sample file, sending the features to a classifier model for training to obtain the prediction model.

3. The identification-based data auto-classification method of claim 2, wherein, After the sample file is acquired, the method further comprises: performing word segmentation, stop word removal, keyword extraction, word matrix conversion, article matrix extraction, and matrix dimension reduction processing on the sample file.

4. The identification-based data auto-classification method of claim 2, wherein, The classifier model comprises one of an SVM model, a decision tree model, a K-nearest neighbor model, or a softmax model.

5. The identification-based data auto-classification method of claim 1, wherein, The keyword retrieval comprises project name retrieval, department name retrieval, or sensitive word retrieval.

6. An identification-based data automatic classification and grading apparatus characterized by comprising: The device comprises: a classification module configured to classify to-be-predicted data by using a prediction model based on a classification label to obtain a classification attribute of data identification; an obtaining module configured to obtain a label-based security level according to the classification attribute and a preset classification grading rule; an establishing module configured to perform keyword retrieval to establish a corresponding relationship between a keyword and a security level; a grading module configured to determine whether the keyword is hit; if yes, taking the security level corresponding to the keyword, subtracting a corresponding preset value corresponding to the label-based security level to obtain a joint grading result; if no, taking the security level in a type rule according to the classification result as a grading result.

Images