Big data supervision method and model for multi-scene interaction

Abstract

The invention belongs to the field of informatization data, and particularly relates to a multi-scene interaction big data supervision method and model, and the method comprises the steps: obtaining data resources in different scenes, carrying out the feature extraction and labeling of the data resources in different scenes, and building a feature library corresponding to different scenes; unifying an API (Application Program Interface) data interface, integrating feature libraries in different scenes to a commemorative inspection data center and a cloud data platform, and carrying out interaction of different scenes in the commemorative inspection data center or the cloud data platform; a plurality of hierarchies from top to bottom are classified into different scenes, the same data feature processing is performed on each scene under a unified AP I interface, a feature library corresponding to each scene is established, a plurality of supervision scenes are established according to requirements, and multi-scene interaction is performed by setting custom association conditions. According to the method, the analysis library which is rich in data and updated in real time in the corresponding supervision scene can be quickly obtained, multi-dimensional supervision is carried out in massive data, and problems and clues are quickly found through multi-data linkage analysis.

Description

Technical Field

[0001] The present invention relates to the field of informationized data, and in particular to a multi-scenario interactive big data supervision method and model. Background Art

[0002] Big data can map interpersonal relationships and connect different data sets through identifiers, which helps to draw inferences from joint data. Big data can also help supervisory agencies gather more "long-tail data", aggregate data scattered in different spaces, form a data synergy, and improve the efficiency of supervision. Practices at multiple levels and in multiple places have shown that the use of big data to carry out discipline inspection and supervision has become a preferred supervision method for discipline inspection and supervision agencies.

[0003] Due to the different levels of development of big data technology, the resources and capabilities of discipline inspection and supervision agencies at all levels to mobilize big data are also different. Some provincial-level discipline inspection and supervision agencies can try to establish data supervision platforms through data aggregation, while the big data supervision methods that grassroots discipline inspection and supervision agencies can use and the big data resources that can be mobilized are generally limited. They can only put forward the idea of ​​"big data supervision" but find it difficult to establish an efficient big data supervision platform. Summary of the invention

[0004] (I) Purpose of the invention

[0005] In order to solve the technical problems existing in the background technology, the present invention proposes a big data supervision method and model for multi-scenario interaction, which has the characteristics of hierarchical setting of multiple scenes, establishing a feature library corresponding to the scenes, and unifying the API data interface for interaction between multiple scenes and supervision scenes, thereby reducing the requirements for supervision technology.

[0006] (II) Technical solution

[0007] In order to solve the above technical problems, the present invention provides a multi-scenario interactive big data supervision method, comprising:

[0008] Acquire data resources in different scenarios, extract and label features of data resources in different scenarios, and establish feature libraries corresponding to different scenarios;

[0009] Unify the API data interface, integrate the feature libraries in different scenarios into the discipline inspection data center and cloud data platform, and interact with different scenarios in the discipline inspection data center or cloud data platform;

[0010] Establish at least one supervision scenario, call the association model in the supervision scenario and customize the association conditions, and interact with the feature libraries corresponding to different scenarios;

[0011] In any supervision scenario, retain the feature library that meets the custom association conditions, call the corresponding data resources according to the feature library, and establish the analysis library corresponding to the current supervision scenario;

[0012] In any supervision scenario, after the called association model changes the custom association conditions, the analysis library established after associating and interacting with the feature libraries of different scenarios is updated in real time.

[0013] Preferably, different scenes are arranged in a top-down hierarchy, including a current-level scene, a lower-level scene, and a secondary-level scene, and data resources are obtained from the current-level, lower-level, and secondary-level scenes for feature extraction and labeling.

[0014] Preferably, feature extraction includes text processing, image processing and big data processing;

[0015] Labels are added to multiple features extracted from data resources that have undergone text processing, image processing, and big data processing, and the features and labels of the data resources are classified and stored to form feature libraries corresponding to different scenarios, namely, the current-level scene feature library, the lower-level scene feature library, and the secondary scene feature library.

[0016] Preferably, the unified API data interface is to create a set of standardized methods and specifications so that clients in different scenarios can exchange data through the same interface.

[0017] Preferably, the discipline inspection data center corresponds to the current level scenario and can uniformly control the lower level scenarios and secondary scenarios;

[0018] The cloud data platform is set up on the cloud server based on cloud computing technology, and the feature libraries corresponding to the current level scenes, lower level scenes and sub-scenes are uploaded to the cloud.

[0019] Preferably, when establishing a supervision scenario, a unique identifier (such as ID) and name are assigned to it, and a supervision scenario description is performed: descriptive information is added to the supervision scenario so that the user can understand the purpose and use of the supervision scenario.

[0020] Preferably, the association model in the supervision scenario is associated with the feature libraries in different scenarios, and when the customized association condition is triggered, the data in the corresponding feature library is obtained.

[0021] Preferably, the analysis library is processed to form a key supervision indicator library through predictive evaluation, expert experience, case reference, historical data analysis and other methods, and the supervision indicators, analysis scope and warning threshold can be flexibly configured, and indicators can be added, edited and deleted.

[0022] Preferably, the updating of the supervision scenario includes: adding a custom association condition and reducing a custom association condition;

[0023] The updates of feature libraries in different scenarios include: daily updates, weekly updates, monthly updates, quarterly updates and annual updates.

[0024] The present invention also provides a multi-scenario interactive big data supervision model, including:

[0025] Feature library building module: obtains data resources in different scenarios, extracts and labels features of data resources in different scenarios, and builds feature libraries corresponding to different scenarios;

[0026] Feature library integration module: Unify the API data interface, integrate feature libraries in different scenarios into the discipline inspection data center and cloud data platform, and interact with different scenarios in the discipline inspection data center or cloud data platform;

[0027] Supervision scenario module: establish at least one supervision scenario, call the association model and customize the association conditions in the supervision scenario, and interact with the feature libraries corresponding to different scenarios;

[0028] Supervision and analysis module: In any supervision scenario, retain the feature library that meets the custom association conditions, call the corresponding data resources according to the feature library, and establish the analysis library corresponding to the current supervision scenario;

[0029] Supervision update module: In any supervision scenario, after the called association model changes the custom association conditions, the analysis library established after associating and interacting with the feature libraries of different scenarios is updated in real time.

[0030] The above-mentioned technical scheme of the present invention has the following beneficial technical effects: multiple levels from top to bottom are classified into different scenes, the same data feature processing is performed on each scene under a unified API interface, a feature library corresponding to each scene is established, and several supervision scenes are established according to needs. Multi-scene interaction is performed by setting custom association conditions, and an analysis library with rich data and real-time updates in the corresponding supervision scene can be quickly obtained, multi-dimensional supervision is performed in massive data, and problems are quickly found and clues are discovered through multi-data linkage analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 Schematic diagram of the big data supervision method of the present invention. DETAILED DESCRIPTION

[0032] In order to make the purpose, technical scheme and advantages of the present invention clearer, the present invention is further described in detail below in conjunction with specific embodiments and with reference to the accompanying drawings. It should be understood that these descriptions are only exemplary and are not intended to limit the scope of the present invention. In addition, in the following description, the description of well-known structures and technologies is omitted to avoid unnecessary confusion of the concept of the present invention.

[0033] like Figure 1 As shown, the present invention proposes a multi-scenario interactive big data supervision method, comprising the following steps:

[0034] S101, acquiring data resources in different scenarios, extracting and labeling features of the data resources in different scenarios, and establishing feature libraries corresponding to different scenarios;

[0035] In S101, different scenarios are set up in a top-down hierarchy, including the current scenario, the lower scenario, and the secondary scenario, and data resources are obtained from the current scenario, the lower scenario, and the secondary scenario to ensure the diversity and comprehensiveness of the data;

[0036] Among them, feature extraction includes text processing, image processing and big data processing;

[0037] Text processing is the extraction of word frequency from data resources: by counting the number of times each word appears in the text, we can understand the main content and keywords of the text, which can be applied to information retrieval, text classification, etc.

[0038] Image processing is the edge detection of data resources: the edge is the place where the grayscale changes the most in the image, and it is also an important part of the image features. Through edge detection algorithms, such as the Canny edge detector, the edge information in the image can be extracted, providing a basis for subsequent image analysis and recognition;

[0039] Big data processing includes: principal component analysis (PCA) and linear discriminant analysis (LDA);

[0040] PCA maps high-dimensional data to low-dimensional space through linear transformation, removes redundant information, and retains the main features of the data, which can extract the most representative features of the data resources;

[0041] LDA uses linear combination to separate the projections of data points of different categories on the combination as much as possible. In feature extraction, LDA can extract the most discriminative features for classification tasks;

[0042] Labels are added to multiple features extracted from data resources that have undergone text processing, image processing, and big data processing, and the features and labels of the data resources are classified and stored to form feature libraries corresponding to different scenarios, namely, the current-level scene feature library, the lower-level scene feature library, and the secondary scene feature library.

[0043] It should be added that the feature libraries for different scenarios are updated regularly, with update times including daily, weekly, monthly, quarterly and annual updates to ensure the timeliness and accuracy of the data.

[0044] S102. Unify the API data interface, integrate the feature libraries under different scenarios into the discipline inspection data center and the cloud data platform, and interact with different scenarios in the discipline inspection data center or the cloud data platform;

[0045] In S102, the unified API data interface is to create a set of standardized methods and specifications so that clients in different scenarios (such as web front-ends, mobile applications, third-party services, etc.) can exchange data through the same interface. The unified API data interface makes the different scenarios maintainable, scalable and interoperable in addition to interaction.

[0046] Among them, the discipline inspection data center corresponds to the current level scene, and can uniformly control the lower level scenes and secondary scenes to avoid the inconvenience of switching multiple scenes. The discipline inspection data center calls the current level scene feature library, the lower level scene feature library, and the secondary scene feature library for interactive supervision;

[0047] The cloud data platform is set up on a cloud server based on cloud computing technology. The feature libraries corresponding to the current level scenes, lower level scenes and secondary scenes are uploaded to the cloud, making it convenient for the interaction between the current level scene feature library, lower level scene feature library and secondary scene feature library in the cloud. The purpose is to take into account the differences in network and information levels in different scenarios, and to take into account both stand-alone and cloud deployment methods. The feature libraries can be connected to the discipline inspection data center through the API data interface, and dynamically displayed through a digital large screen, so that the lower level scenes and secondary scenes with lower network and information levels can interact with each other for supervision.

[0048] It should be added that: by setting a unified API data interface, application scenarios other than the current level scenario, lower level scenario and secondary scenario can be expanded;

[0049] For example, in addition to being used by the Discipline Inspection Data Center itself, the unified API data interface can also be shared with sister unit scenarios or inspection scenarios, expanding the feature library corresponding to sister unit scenarios and inspection scenarios, conducting multi-scenario interactions, and increasing the scope of application of supervision.

[0050] S103, establishing at least one supervision scenario, calling the association model in the supervision scenario and defining the association conditions, and performing association interaction with the feature libraries corresponding to different scenarios;

[0051] When establishing a supervision scenario, assign it a unique identifier (such as an ID) and a name, and describe the supervision scenario: add descriptive information to the supervision scenario so that users can understand the purpose and use of the supervision scenario;

[0052] Calling the association model: The association model is trained on big data in terms of decision trees, support vector machines, and neural networks;

[0053] Custom association conditions: Allow users to set custom association conditions for association models. These conditions can be based on logical expressions of data fields, such as "Field A is equal to a certain value" or "Field B is within a certain range."

[0054] In S101, feature libraries are established for different scenarios. The feature library is a data set containing the features required for a specific scenario. According to the custom association conditions, the association model in the supervision scenario is associated with the feature libraries in different scenarios. When the custom association conditions are triggered, the data in the corresponding feature library is obtained.

[0055] It can be understood that: when a supervision scenario is activated, the association model queries data from different scenario feature libraries according to the customized association conditions, and performs corresponding subsequent operations such as data verification, statistical analysis, anomaly detection, etc. according to the output of the association model.

[0056] In one embodiment, multiple supervision scenarios are set, and in each supervision scenario, the association model has different customized association conditions, that is, different supervision scenarios have different association interaction tasks.

[0057] S104: In any supervision scenario, retain the feature library that meets the custom association conditions, call the corresponding data resources according to the feature library, and establish an analysis library corresponding to the current supervision scenario;

[0058] Since the feature library that meets the custom association conditions only has the feature parts that have been processed by text, image and big data, the initial data resources corresponding to the feature library before processing are called, and the initial data resources are processed and combined to establish an analysis library corresponding to the current supervision scenario.

[0059] Process the analysis database and form a key supervision indicator database through methods such as pre-judgment evaluation, expert experience, case reference, and historical data analysis. The supervision indicators, analysis scope, and warning threshold can be flexibly configured, and indicators can be added, edited, and deleted.

[0060] With supervision, management and control as the focus, data analysis is carried out in a multi-dimensional and refined manner. The analysis results can be presented and the analysis data can be exported in lists, pie charts, bar charts, radar charts, trend charts, line charts, etc.

[0061] S105. In any supervision scenario, after the called association model changes the custom association condition, the analysis library established after the association interaction with the feature libraries of different scenarios is updated in real time.

[0062] Among them, when the association model called in the supervision scenario changes the custom association condition:

[0063] Add custom association conditions: re-interact the original associated scene feature libraries, filter out the parts that do not meet the added custom association conditions, and update the analysis library;

[0064] Reduce custom association conditions: re-interact with the feature libraries that do not meet the custom association conditions during the initial interaction, add some feature libraries that meet the second custom association conditions into the analysis library, and update the analysis library;

[0065] The custom association conditions remain unchanged. When feature libraries are added in different scenarios, the added feature libraries are associated and interacted, and the feature libraries that meet the custom association conditions are added to the analysis library to update the analysis library.

[0066] In this embodiment, multiple levels from top to bottom are classified into different scenarios, and the same data feature processing is performed on each scenario under a unified API interface. A feature library corresponding to each scenario is established, and several supervision scenarios are established according to needs. Multi-scenario interaction is performed by setting custom association conditions. An analysis library with rich data and real-time updates in the corresponding supervision scenario can be quickly obtained, multi-dimensional supervision is performed in massive data, and problems and clues are quickly found through multi-data linkage analysis.

[0067] A multi-scenario interactive big data supervision model, including:

[0068] Feature library building module: obtains data resources in different scenarios, extracts and labels features of data resources in different scenarios, and builds feature libraries corresponding to different scenarios;

[0069] Feature library integration module: Unify the API data interface, integrate feature libraries in different scenarios into the discipline inspection data center and cloud data platform, and interact with different scenarios in the discipline inspection data center or cloud data platform;

[0070] Supervision scenario module: establish at least one supervision scenario, call the association model and customize the association conditions in the supervision scenario, and interact with the feature libraries corresponding to different scenarios;

[0071] Supervision and analysis module: In any supervision scenario, retain the feature library that meets the custom association conditions, call the corresponding data resources according to the feature library, and establish the analysis library corresponding to the current supervision scenario;

[0072] Supervision update module: In any supervision scenario, after the called association model changes the custom association conditions, the analysis library established after associating and interacting with the feature libraries of different scenarios is updated in real time.

[0073] It should be understood that the above specific embodiments of the present invention are only used to illustrate or explain the principles of the present invention, and do not constitute a limitation of the present invention. Therefore, any modifications, equivalent substitutions, improvements, etc. made without departing from the spirit and scope of the present invention should be included in the protection scope of the present invention. In addition, the appended claims of the present invention are intended to cover all changes and modifications that fall within the scope and boundaries of the appended claims, or the equivalent forms of such scope and boundaries.

Claims

1. A multi-scenario interactive big data supervision method, characterized in that: include: Acquire data resources in different scenarios, extract and label features of data resources in different scenarios, and establish feature libraries corresponding to different scenarios; Unify the API data interface, integrate the feature libraries under different scenarios into the discipline inspection data center and cloud data platform, and interact with different scenarios in the discipline inspection data center or cloud data platform; Establish at least one supervision scenario, call the association model in the supervision scenario and customize the association conditions, and interact with the feature libraries corresponding to different scenarios; In any supervision scenario, retain the feature library that meets the custom association conditions, call the corresponding data resources according to the feature library, and establish the analysis library corresponding to the current supervision scenario; In any supervision scenario, after the called association model changes the custom association conditions, the analysis library established after associating and interacting with the feature libraries of different scenarios is updated in real time.

2. According to the multi-scenario interactive big data supervision method of claim 1, it is characterized in that: Different scenes are set up in a top-down hierarchy, including the current-level scene, the lower-level scene, and the secondary scene. Data resources are obtained from the current-level, lower-level, and secondary scenes for feature extraction and labeling.

3. According to the multi-scenario interactive big data supervision method of claim 1, it is characterized in that: Feature extraction includes text processing, image processing and big data processing; Labels are added to multiple features extracted from data resources that have undergone text processing, image processing, and big data processing, and the features and labels of the data resources are classified and stored to form feature libraries corresponding to different scenarios, namely, the current-level scene feature library, the lower-level scene feature library, and the secondary scene feature library.

4. According to the multi-scenario interactive big data supervision method of claim 1, it is characterized in that: The unified API data interface creates a set of standardized methods and specifications so that clients in different scenarios can exchange data through the same interface.

5. The multi-scenario interactive big data supervision method according to claim 1 is characterized in that: The Discipline Inspection Data Center corresponds to the current level scenario and can conduct unified management and control of lower-level scenarios and secondary scenarios; The cloud data platform is set up on the cloud server based on cloud computing technology, and the feature libraries corresponding to the current level scenes, lower level scenes and sub-scenes are uploaded to the cloud.

6. The multi-scenario interactive big data supervision method according to claim 1, characterized in that: When establishing a supervision scenario, assign it a unique identifier (such as an ID) and a name, and describe the supervision scenario: add descriptive information to the supervision scenario so that users can understand the purpose and use of the supervision scenario.

7. The multi-scenario interactive big data supervision method according to claim 1, characterized in that: In the supervision scenario, the association model is associated with the feature libraries in different scenarios. When the custom association condition is triggered, the data in the corresponding feature library is obtained.

8. The multi-scenario interactive big data supervision method according to claim 1, characterized in that: The analysis library is processed to form a key supervision indicator library through predictive evaluation, expert experience, case reference, historical data analysis and other methods. The supervision indicators, analysis scope and warning thresholds can be flexibly configured, and indicators can be added, edited and deleted.

9. The multi-scenario interactive big data supervision method according to claim 1, characterized in that: Updates to supervision scenarios include: adding custom association conditions and reducing custom association conditions; The updates of feature libraries in different scenarios include: daily updates, weekly updates, monthly updates, quarterly updates and annual updates.

10. A multi-scenario interactive big data supervision model, characterized in that: include: Feature library building module: obtains data resources in different scenarios, extracts and labels features of data resources in different scenarios, and builds feature libraries corresponding to different scenarios; Feature library integration module: Unify the API data interface, integrate feature libraries in different scenarios into the discipline inspection data center and cloud data platform, and interact with different scenarios in the discipline inspection data center or cloud data platform; Supervision scenario module: establish at least one supervision scenario, call the association model and customize the association conditions in the supervision scenario, and interact with the feature libraries corresponding to different scenarios; Supervision and analysis module: In any supervision scenario, retain the feature library that meets the custom association conditions, call the corresponding data resources according to the feature library, and establish the analysis library corresponding to the current supervision scenario; Supervision update module: In any supervision scenario, after the called association model changes the custom association conditions, the analysis library established after associating and interacting with the feature libraries of different scenarios is updated in real time.

Images