A monitoring data screening method and system

Abstract

The present invention provides a monitoring data screening method and system, the method comprising: obtaining a plurality of decision attribute values ​​from a decision mechanism based on decision attributes, and assigning a first ID number to form a first data pair, thereby obtaining a first equivalent set; obtaining monitoring values ​​from a monitoring mechanism including a plurality of monitoring attributes, and assigning a second ID number to form a second data pair, thereby obtaining a second equivalent set; obtaining decision information entropy through the first equivalent set; determining to obtain monitoring condition entropy based on the first equivalent set and the second equivalent set; determining a gain value of the monitoring attribute based on the decision information entropy and the monitoring condition entropy, and judging whether the monitoring attribute needs to be filtered through the gain value. Based on the above method, the correlation degree between the monitoring attribute and the decision record is analyzed, the monitoring attribute with a lower correlation degree is screened out, a large amount of redundant data and invalid data are eliminated, and the monitoring efficiency and monitoring accuracy are improved.

Description

Technical Field

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

[0002] In the scenario where multiple agencies jointly monitor and acquire data, how to complete data acquisition from multiple agencies while strictly ensuring data privacy and security has become a key technical challenge that needs to be overcome urgently.

[0003] In recent years, federated learning, as an emerging data processing technology, has provided a new solution for joint security monitoring by multiple institutions. Federated learning allows multiple institutions to process and analyze data locally, and only upload the encrypted processing results to the coordinating party for aggregation and analysis, thereby avoiding the risk of data leakage.

[0004] However, based on the framework of federated learning, how to efficiently and accurately screen the data sources obtained after monitoring by multiple institutions is still a technical challenge. The data sources of multiple institutions are extensive and diverse, and there is a large amount of redundant and invalid data, which will lead to a decrease in the overall monitoring efficiency and accuracy. Summary of the invention

[0005] In view of the shortcomings of the prior art, the purpose of the present invention is to provide a monitoring data screening method and system, aiming to solve the technical problems in the prior art that the data sources of multiple institutions are wide and diverse, and there are a large amount of redundant and invalid data, which will lead to a decrease in the overall monitoring efficiency and accuracy.

[0006] In order to achieve the above-mentioned purpose, in a first aspect, an embodiment of the present application provides a monitoring data screening method for filtering monitoring attributes of a plurality of monitoring agencies, the monitoring data screening method comprising the following steps:

[0007] Based on the decision attribute, a plurality of decision attribute values ​​corresponding to the monitoring object are obtained from the decision record of the decision-making body, a first ID number is assigned to each of the decision attribute values, the first ID number and the decision attribute value are combined into a first data pair, and a plurality of the first data pairs are equivalently divided to form a first equivalence set including a plurality of first equivalence classes;

[0008] Acquire a monitoring record corresponding to a monitoring object from a monitoring mechanism, wherein the monitoring mechanism includes a plurality of monitoring attributes, extract a monitoring value from the monitoring record based on the monitoring attribute, assign a second ID number to each of the monitoring values, combine the second ID number with the monitoring value into a second data pair, perform equivalence partitioning on a plurality of the second data pairs to form a second equivalence set including a plurality of second equivalence classes, wherein the number of the second equivalence sets is the same as the number of the monitoring attributes;

[0009] Acquire the decision information entropy corresponding to the decision attribute through the first equivalent set;

[0010] Determine, based on the first equivalence set and the second equivalence set, a monitoring condition entropy of a monitoring attribute corresponding to the second equivalence set;

[0011] A gain value of the monitoring attribute is determined based on the decision information entropy and the monitoring condition entropy, and whether the monitoring attribute needs to be filtered is determined by the gain value.

[0012] Compared with the prior art, the beneficial effects of the present invention are: only the first equivalent set including the first ID number and the second equivalent set including the second ID number are processed, thereby ensuring the security of the data; through the information entropy theory, the uncertainty of the data is calculated to measure the information content and redundancy of the data, that is, by obtaining the decision information entropy and the monitoring condition entropy, the correlation between the monitoring attribute and the decision record is analyzed, and then the monitoring attributes with a lower correlation are screened out from a large number of monitoring attributes, avoiding the monitoring values ​​of the monitoring attributes of this part from being mixed with important monitoring values, eliminating more redundant data and invalid data, and improving the monitoring efficiency and monitoring accuracy.

[0013] Further, the step of performing equivalence partitioning on the plurality of first data pairs to form a first equivalence set including a plurality of first equivalence classes includes:

[0014] comparing the decision attribute values ​​in different first data pairs;

[0015] If the same decision attribute value exists, the first ID numbers corresponding to the same decision attribute value are merged into a first set, and the remaining first ID numbers are respectively determined as second sets;

[0016] The first set and the second set are both defined as a first equivalence class, and a number of the first equivalence classes are combined into a first equivalence set.

[0017] Furthermore, the step of obtaining the decision information entropy corresponding to the decision attribute through the first equivalent set is specifically:

[0018] Obtaining a probability value of the first equivalence class in the first equivalence set;

[0019] The decision information entropy corresponding to the decision attribute is calculated based on the probability value of the first equivalence class.

[0020] Furthermore, the calculation formula of the probability value of the first equivalence class is:

[0021] ,

[0022] in, represents the probability value of the i-th first equivalence class, represents the number of first ID numbers in the i-th first equivalence class, Indicates the number of first ID numbers in the first equivalence set.

[0023] Furthermore, the calculation formula of the decision information entropy is:

[0024] ,

[0025] in, represents the decision information entropy, represents the probability value of the i-th first equivalence class, represents the number of first equivalence classes, Represents a logarithmic function.

[0026] Furthermore, the step of determining the monitoring condition entropy of the monitoring attribute corresponding to the second equivalence set based on the first equivalence set and the second equivalence set includes:

[0027] In a second equivalence set corresponding to a monitoring attribute, taking the second equivalence class as a condition and taking the first equivalence set as a result, determining a single conditional entropy of the second equivalence class;

[0028] The monitoring condition entropy of the monitoring attribute is determined by using a plurality of the single condition entropies.

[0029] Furthermore, the calculation formula of the single conditional entropy is:

[0030] ,

[0031] in, represents the single conditional entropy of the qth second equivalence class corresponding to the jth monitoring attribute under the ith monitoring agency, It represents the probability of taking the qth second equivalence class corresponding to the jth monitoring attribute under the ith monitoring agency as the conditional value and the nth first equivalence class as the result value, represents the logarithmic function, Represents the number of first equivalence classes.

[0032] Furthermore, the calculation formula of the monitoring condition entropy is:

[0033] ,

[0034] in, represents the monitoring condition entropy of the jth monitoring attribute of the i-th monitoring agency, represents the single conditional entropy of the mth second equivalence class corresponding to the jth monitoring attribute under the i-th monitoring agency, represents the probability value of the mth second equivalence class corresponding to the jth monitoring attribute under the i-th monitoring agency, It represents the number of second equivalence classes corresponding to the jth monitoring attribute under the i-th monitoring organization.

[0035] Furthermore, the step of determining whether to filter the monitoring attribute according to the gain value includes:

[0036] comparing the gain value with an attribute threshold;

[0037] If the gain value is greater than the attribute threshold, retaining the monitoring attribute corresponding to the gain value greater than the attribute threshold;

[0038] If the gain value is smaller than the attribute threshold, the monitoring attribute corresponding to the gain value smaller than the attribute threshold is eliminated.

[0039] In a second aspect, an embodiment of the present application provides a monitoring data screening system, which is applied to the monitoring data screening method as described in the first aspect above, and the system includes:

[0040] A first acquisition module is used to acquire a plurality of decision attribute values ​​corresponding to the monitoring object from the decision record of the decision-making body based on the decision attribute, assign a first ID number to each of the decision attribute values, combine the first ID number and the decision attribute value into a first data pair, and perform equivalent division on a plurality of the first data pairs to form a first equivalent set including a plurality of first equivalent classes;

[0041] A second acquisition module is used to acquire a monitoring record corresponding to a monitoring object from a monitoring mechanism, wherein the monitoring mechanism includes a plurality of monitoring attributes, extract monitoring values ​​from the monitoring records based on the monitoring attributes, assign a second ID number to each of the monitoring values, combine the second ID number with the monitoring value into a second data pair, perform equivalence partitioning on a plurality of the second data pairs to form a second equivalence set including a plurality of second equivalence classes, and the number of the second equivalence sets is the same as the number of the monitoring attributes;

[0042] A first analysis module, configured to obtain a decision information entropy corresponding to the decision attribute through the first equivalent set;

[0043] A second analysis module, configured to determine, based on the first equivalence set and the second equivalence set, a monitoring condition entropy of a monitoring attribute corresponding to the second equivalence set;

[0044] A judgment module is used to determine a gain value of the monitoring attribute based on the decision information entropy and the monitoring condition entropy, and to judge whether the monitoring attribute needs to be filtered according to the gain value.

[0045] In a third aspect, an embodiment of the present application provides a computer, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the monitoring data screening method as described in the first aspect above is implemented.

[0046] In a fourth aspect, an embodiment of the present application provides a storage medium on which a computer program is stored. When the computer program is executed by a processor, the monitoring data screening method as described in the first aspect above is implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 Flow chart of the monitoring data screening method in the first embodiment of the present invention;

[0048] Figure 2 It is a structural block diagram of a monitoring data screening system in a second embodiment of the present invention;

[0049] The following specific implementation manner will further illustrate the present invention in conjunction with the above-mentioned drawings. DETAILED DESCRIPTION

[0050] In order to facilitate the understanding of the present invention, the present invention will be described more fully below with reference to the relevant drawings. Several embodiments of the present invention are given in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive.

[0051] It should be noted that when an element is referred to as being "fixed to" another element, it may be directly on the other element or there may be a central element. When an element is considered to be "connected to" another element, it may be directly connected to the other element or there may be a central element at the same time. The terms "vertical", "horizontal", "left", "right" and similar expressions used herein are for illustrative purposes only.

[0052] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which the present invention belongs. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0053] See also Figure 1The monitoring data screening method provided by the first embodiment of the present invention is used to filter the monitoring attributes of several monitoring agencies to extract data according to the filtered monitoring attributes. The monitoring data screening method includes the following steps:

[0054] S10: Based on the decision attribute, a plurality of decision attribute values ​​corresponding to the monitoring object are obtained from the decision record of the decision-making body, a first ID number is assigned to each of the decision attribute values, the first ID number and the decision attribute value are combined into a first data pair, and a plurality of the first data pairs are equivalently divided to form a first equivalence set including a plurality of first equivalence classes;

[0055] In this embodiment, the monitoring object is weather, the decision-making body is a weather forecast center, the plurality of decision data records are a plurality of weather records, the decision attribute is weather, and weather is used as a keyword to extract weather conditions from weather records.

[0056] The step S10 comprises:

[0057] S110: comparing the decision attribute values ​​in different first data pairs;

[0058] S120: If the same decision attribute value exists, the first ID numbers corresponding to the same decision attribute value are merged into a first set, and the remaining first ID numbers are respectively determined as second sets;

[0059] S130: defining the first set and the second set as a first equivalence class, and combining a plurality of the first equivalence classes into a first equivalence set;

[0060] Taking 3 days of weather records as an example, if the decision attribute values ​​are sunny, cloudy, and sunny, then after the first ID number is assigned, the three first data pairs are 1, sunny; 2, cloudy; and 3, sunny. At this time, there is the same decision attribute value. Then the first set is (1, 3), the second set is (2), and the first equivalent set is {(1, 3), (2)}.

[0061] It can be understood that if there is no identical decision attribute value, the first ID numbers corresponding to different decision attribute values ​​are respectively divided into a plurality of third sets;

[0062] Taking the weather records of three days as an example, if the decision attribute values ​​are sunny, cloudy, and rainy, then after the first ID number is assigned, the three first data pairs are 1, sunny; 2, cloudy; 3, rainy. At this time, there is no identical decision attribute value, then the three third sets are (1), (2), (3), and the first equivalent set is {(1), (2), (3)}. It should be noted that the acquisition of the first equivalent set can be completed within the decision-making body, and then the first equivalent set is transmitted to the central server for subsequent processing to ensure the security and privacy of the data.

[0063] S20: Acquire a monitoring record corresponding to the monitoring object from a monitoring mechanism, wherein the monitoring mechanism includes a plurality of monitoring attributes, extract a monitoring value from the monitoring record based on the monitoring attribute, assign a second ID number to each of the monitoring values, combine the second ID number with the monitoring value into a second data pair, perform equivalence partitioning on a plurality of the second data pairs to form a second equivalence set including a plurality of second equivalence classes, wherein the number of the second equivalence sets is the same as the number of the monitoring attributes;

[0064] In this embodiment, the monitoring mechanism is a base station. Different monitoring mechanisms may collect the same or different data. When the monitoring object is weather, if there are three monitoring mechanisms, the first monitoring mechanism may collect air pressure, wind speed, and temperature, the second monitoring mechanism may collect humidity and wind speed, and the third monitoring mechanism may collect wind speed and temperature. Then the first monitoring mechanism includes three monitoring attributes, namely, air pressure, wind speed, and temperature. The monitoring attributes of the remaining monitoring mechanisms are similar.

[0065] It can be understood that when the monitoring object is weather, the monitoring record is a data collection record of the monitoring agency for 3 days. From the monitoring record, a corresponding match is performed based on the monitoring attribute to obtain a monitoring value corresponding to the monitoring attribute.

[0066] The acquisition method of the second equivalence set is consistent with that of the first equivalence set, which will not be elaborated here. Taking the air pressure of the first monitoring agency as an example, in the 3-day monitoring record, the monitoring values ​​corresponding to the air pressure in the first monitoring agency are high, medium and low respectively. After assigning the second ID number, the 3 second data pairs are 1, high; 2, medium; 3, low, then the second equivalence class is: (1), (2), (3), then the second equivalent combination corresponding to the air pressure is {(1), (2), (3)}. It can be understood that if there are 3 monitoring attributes in the first monitoring agency, there are 3 corresponding second equivalence sets. It should be noted that the acquisition of the second equivalence set can be completed within the monitoring agency, and then the second equivalence set is transmitted to the central server for subsequent processing to ensure the security and privacy of the data.

[0067] S30: Acquire decision information entropy corresponding to the decision attribute through the first equivalent set;

[0068] Specifically, the step S30 includes:

[0069] S310: Obtain a probability value of the first equivalence class in the first equivalence set;

[0070] The calculation formula of the probability value of the first equivalence class is:

[0071] ,

[0072] in, represents the probability value of the i-th first equivalence class, represents the number of first ID numbers in the i-th first equivalence class, Indicates the number of first ID numbers in the first equivalence set.

[0073] S320: Calculating the decision information entropy corresponding to the decision attribute based on the probability value of the first equivalence class;

[0074] The calculation formula of the decision information entropy is:

[0075] ,

[0076] in, represents the decision information entropy, represents the probability value of the i-th first equivalence class, represents the number of first equivalence classes, Represents a logarithmic function.

[0077] S40: Determine, based on the first equivalence set and the second equivalence set, a monitoring condition entropy of a monitoring attribute corresponding to the second equivalence set;

[0078] The step S40 comprises:

[0079] S410: In a second equivalence set corresponding to a certain monitoring attribute, determine a single conditional entropy of the second equivalence class with the second equivalence class as a condition and the first equivalence set as a result;

[0080] The calculation formula of the single conditional entropy is:

[0081] ,

[0082] in, represents the single conditional entropy of the qth second equivalence class corresponding to the jth monitoring attribute under the ith monitoring agency, It represents the probability of taking the qth second equivalence class corresponding to the jth monitoring attribute under the ith monitoring agency as the conditional value and the nth first equivalence class as the result value, represents the logarithmic function, Represents the number of first equivalence classes.

[0083] Taking the example that there are three second equivalence classes in the first monitoring mechanism, it is necessary to calculate the single conditional entropy of the three second equivalence classes respectively.

[0084] S420: Determine the monitoring condition entropy of the monitoring attribute through a plurality of the single condition entropies;

[0085] The calculation formula of the monitoring condition entropy is:

[0086] ,

[0087] in, represents the monitoring condition entropy of the jth monitoring attribute of the i-th monitoring agency, represents the single conditional entropy of the mth second equivalence class corresponding to the jth monitoring attribute under the i-th monitoring agency, represents the probability value of the mth second equivalence class corresponding to the jth monitoring attribute under the i-th monitoring agency, It represents the number of second equivalence classes corresponding to the jth monitoring attribute under the i-th monitoring organization.

[0088] S50: determining a gain value of the monitoring attribute based on the decision information entropy and the monitoring condition entropy, and judging whether the monitoring attribute needs to be filtered according to the gain value;

[0089] The gain value can be obtained by subtracting the monitoring condition entropy from the decision information entropy. The gain value represents the degree of association between a certain monitoring attribute under a certain monitoring mechanism and the decision record.

[0090] The step S50 comprises:

[0091] S510: Compare the gain value with the attribute threshold;

[0092] S520: If the gain value is greater than the attribute threshold, retaining the monitoring attribute corresponding to the gain value greater than the attribute threshold;

[0093] S530: If the gain value is less than the attribute threshold, remove the monitoring attribute corresponding to the gain value less than the attribute threshold;

[0094] Furthermore, after completing the comparison between the gain value and the attribute threshold, the monitoring mechanism, the monitoring attribute under the monitoring mechanism, the gain value of the monitoring attribute, and the comparison result between the gain value and the attribute threshold are recorded in a table, and the table is transmitted to the decision-making mechanism, and the decision-making mechanism can filter the monitoring attributes under different monitoring mechanisms according to the comparison result, and when extracting data, the monitoring values ​​corresponding to the monitoring attributes with gain values ​​greater than the attribute threshold are extracted, and the monitoring values ​​corresponding to the monitoring attributes with gain values ​​greater than the attribute threshold are not extracted. When extracting the monitoring values ​​corresponding to the monitoring attributes with gain values ​​greater than the attribute threshold, the monitoring values ​​corresponding to the monitoring attributes with larger gain values ​​can be preferentially extracted according to the size of the gain values.

[0095] Only the first equivalent set including the first ID number and the second equivalent set including the second ID number are processed to ensure the security of the data; through the information entropy theory, the uncertainty of the data is calculated to measure the information content and redundancy of the data, that is, by obtaining the decision information entropy and the monitoring condition entropy, the correlation between the monitoring attribute and the decision record is analyzed, and then the monitoring attributes with a lower correlation are screened out from the numerous monitoring attributes to avoid the monitoring values ​​of the monitoring attributes being mixed with the important monitoring values, thereby eliminating more redundant data and invalid data and improving the monitoring efficiency and accuracy.

[0096] See also Figure 2 The second embodiment of the present invention provides a monitoring data screening system, which is applied to the monitoring data screening method described in the above embodiment, and will not be repeated here. As used below, the terms "module", "unit", "subunit", etc. can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0097] The system comprises:

[0098] A first acquisition module 10 is used to acquire a plurality of decision attribute values ​​corresponding to the monitoring object from the decision record of the decision-making body based on the decision attribute, assign a first ID number to each of the decision attribute values, combine the first ID number and the decision attribute value into a first data pair, and perform equivalent division on a plurality of the first data pairs to form a first equivalence set including a plurality of first equivalence classes;

[0099] The first acquisition module 10 includes:

[0100] A first unit, configured to compare the decision attribute values ​​in different first data pairs;

[0101] The second unit is used for merging the first ID numbers corresponding to the same decision attribute value into a first set if there are the same decision attribute values, and determining the remaining first ID numbers as second sets respectively;

[0102] A third unit is used to define the first set and the second set as a first equivalence class, and combine a plurality of the first equivalence classes into a first equivalence set;

[0103] A second acquisition module 20 is used to acquire a monitoring record corresponding to a monitoring object from a monitoring mechanism, wherein the monitoring mechanism includes a plurality of monitoring attributes, extract a monitoring value from the monitoring record based on the monitoring attribute, assign a second ID number to each of the monitoring values, combine the second ID number with the monitoring value into a second data pair, perform equivalence partitioning on a plurality of the second data pairs to form a second equivalence set including a plurality of second equivalence classes, and the number of the second equivalence sets is the same as the number of the monitoring attributes;

[0104] A first analysis module 30, configured to obtain a decision information entropy corresponding to the decision attribute through the first equivalent set;

[0105] The first analysis module 30 includes:

[0106] A fourth unit is used to obtain a probability value of the first equivalence class in the first equivalence set;

[0107] A fifth unit, configured to calculate a decision information entropy corresponding to the decision attribute based on the probability value of the first equivalence class;

[0108] A second analysis module 40, configured to determine, based on the first equivalence set and the second equivalence set, a monitoring condition entropy of a monitoring attribute corresponding to the second equivalence set;

[0109] The second analysis module 40 includes:

[0110] A sixth unit is used to determine, in a second equivalence set corresponding to a certain monitoring attribute, a single conditional entropy of the second equivalence class with the second equivalence class as a condition and the first equivalence set as a result;

[0111] A seventh unit, configured to determine a monitoring condition entropy of the monitoring attribute through a plurality of the single condition entropies;

[0112] A judgment module 50, configured to determine a gain value of the monitoring attribute based on the decision information entropy and the monitoring condition entropy, and to judge whether the monitoring attribute needs to be filtered according to the gain value;

[0113] The judging module 50 comprises:

[0114] An eighth unit, configured to compare the gain value with an attribute threshold;

[0115] A ninth unit, configured to retain a monitoring attribute corresponding to the gain value greater than the attribute threshold if the gain value is greater than the attribute threshold;

[0116] The tenth unit is configured to eliminate the monitoring attribute corresponding to the gain value smaller than the attribute threshold if the gain value is smaller than the attribute threshold.

[0117] The present invention also provides a computer, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the monitoring data screening method as described in the above technical solution when executing the computer program.

[0118] The present invention also provides a storage medium on which a computer program is stored. When the computer program is executed by a processor, the monitoring data screening method as described in the above technical solution is implemented.

[0119] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0120] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A monitoring data screening method for filtering monitoring attributes of several monitoring agencies, characterized in that: The monitoring data screening method comprises the following steps: Based on the decision attribute, a plurality of decision attribute values ​​corresponding to the monitoring object are obtained from the decision record of the decision-making body, a first ID number is assigned to each of the decision attribute values, the first ID number and the decision attribute value are combined into a first data pair, and a plurality of the first data pairs are equivalently divided to form a first equivalence set including a plurality of first equivalence classes; Acquire a monitoring record corresponding to a monitoring object from a monitoring mechanism, wherein the monitoring mechanism includes a plurality of monitoring attributes, extract a monitoring value from the monitoring record based on the monitoring attribute, assign a second ID number to each of the monitoring values, combine the second ID number with the monitoring value into a second data pair, perform equivalence partitioning on a plurality of the second data pairs to form a second equivalence set including a plurality of second equivalence classes, wherein the number of the second equivalence sets is the same as the number of the monitoring attributes; Acquire the decision information entropy corresponding to the decision attribute through the first equivalent set; Determine, based on the first equivalence set and the second equivalence set, a monitoring condition entropy of a monitoring attribute corresponding to the second equivalence set; A gain value of the monitoring attribute is determined based on the decision information entropy and the monitoring condition entropy, and whether the monitoring attribute needs to be filtered is determined by the gain value.

2. The monitoring data screening method according to claim 1, characterized in that: The step of performing equivalence partitioning on the plurality of first data pairs to form a first equivalence set including a plurality of first equivalence classes comprises: comparing the decision attribute values ​​in different first data pairs; If the same decision attribute value exists, the first ID numbers corresponding to the same decision attribute value are merged into a first set, and the remaining first ID numbers are respectively determined as second sets; The first set and the second set are both defined as a first equivalence class, and a number of the first equivalence classes are combined into a first equivalence set.

3. The monitoring data screening method according to claim 1, characterized in that: The step of obtaining the decision information entropy corresponding to the decision attribute through the first equivalent set is specifically: Obtaining a probability value of the first equivalence class in the first equivalence set; The decision information entropy corresponding to the decision attribute is calculated based on the probability value of the first equivalence class.

4. The monitoring data screening method according to claim 3, characterized in that: The calculation formula of the probability value of the first equivalence class is: ， in, represents the probability value of the i-th first equivalence class, represents the number of first ID numbers in the i-th first equivalence class, Indicates the number of first ID numbers in the first equivalence set.

5. The monitoring data screening method according to claim 3, characterized in that: The calculation formula of the decision information entropy is: ， in, represents the decision information entropy, represents the probability value of the i-th first equivalence class, represents the number of first equivalence classes, Represents a logarithmic function.

6. The monitoring data screening method according to claim 1, characterized in that: The step of determining the monitoring condition entropy of the monitoring attribute corresponding to the second equivalence set based on the first equivalence set and the second equivalence set comprises: In a second equivalence set corresponding to a monitoring attribute, taking the second equivalence class as a condition and taking the first equivalence set as a result, determining a single conditional entropy of the second equivalence class; The monitoring condition entropy of the monitoring attribute is determined by using a plurality of the single condition entropies.

7. The monitoring data screening method according to claim 6, characterized in that: The calculation formula of the single conditional entropy is: ， in, represents the single conditional entropy of the qth second equivalence class corresponding to the jth monitoring attribute under the ith monitoring agency, It represents the probability of taking the qth second equivalence class corresponding to the jth monitoring attribute under the ith monitoring agency as the conditional value and the nth first equivalence class as the result value, represents the logarithmic function, Represents the number of first equivalence classes.

8. The monitoring data screening method according to claim 6, characterized in that: The calculation formula of the monitoring condition entropy is: ， in, represents the monitoring condition entropy of the jth monitoring attribute of the i-th monitoring agency, represents the single conditional entropy of the mth second equivalence class corresponding to the jth monitoring attribute under the i-th monitoring agency, represents the probability value of the mth second equivalence class corresponding to the jth monitoring attribute under the i-th monitoring agency, It represents the number of second equivalence classes corresponding to the jth monitoring attribute under the i-th monitoring organization.

9. The monitoring data screening method according to claim 1, characterized in that: The step of judging whether to filter the monitoring attribute by the gain value comprises: comparing the gain value with an attribute threshold; If the gain value is greater than the attribute threshold, retaining the monitoring attribute corresponding to the gain value greater than the attribute threshold; If the gain value is smaller than the attribute threshold, the monitoring attribute corresponding to the gain value smaller than the attribute threshold is eliminated.

10. A monitoring data screening system, applied to the monitoring data screening method according to any one of claims 1 to 9, characterized in that: The system comprises: A first acquisition module is used to acquire a plurality of decision attribute values ​​corresponding to the monitoring object from the decision record of the decision-making body based on the decision attribute, assign a first ID number to each of the decision attribute values, combine the first ID number and the decision attribute value into a first data pair, and perform equivalent division on a plurality of the first data pairs to form a first equivalent set including a plurality of first equivalent classes; A second acquisition module is used to acquire a monitoring record corresponding to a monitoring object from a monitoring mechanism, wherein the monitoring mechanism includes a plurality of monitoring attributes, extract monitoring values ​​from the monitoring records based on the monitoring attributes, assign a second ID number to each of the monitoring values, combine the second ID number with the monitoring value into a second data pair, perform equivalence partitioning on a plurality of the second data pairs to form a second equivalence set including a plurality of second equivalence classes, and the number of the second equivalence sets is the same as the number of the monitoring attributes; A first analysis module, configured to obtain a decision information entropy corresponding to the decision attribute through the first equivalent set; A second analysis module, configured to determine, based on the first equivalence set and the second equivalence set, a monitoring condition entropy of a monitoring attribute corresponding to the second equivalence set; A judgment module is used to determine a gain value of the monitoring attribute based on the decision information entropy and the monitoring condition entropy, and to judge whether the monitoring attribute needs to be filtered according to the gain value.

Images