AMC Excess Cause Mining and Analysis Method and System Based on Active Learning Framework

Abstract

The present invention relates to the technical field of analysis of causes of excessive AMC, and specifically discloses a method and system for mining and analyzing causes of excessive AMC using an active learning framework. The method comprises extracting a submatrix sequence corresponding to the AMC monitor from a data matrix sequence when any AMC concentration reaches a preset threshold; marking row and column positions in the submatrix sequence where the amount of data information is less than a preset information threshold, deleting the data at the marked row and column positions in the submatrix sequence to obtain a simplified matrix sequence; using the obtained simplified matrix sequence as a feature and the AMC concentration as a label to construct a sample set, and determining the cause of the excessive AMC based on the sample set; the present invention provides a data volume-based data screening scheme for actively screening the sample set, reducing the number of samples in the sample set, and improving the quality of the samples in the sample set. This preliminary work belongs to the active learning framework and is integrated with the existing recognition process, greatly improving the efficiency of cause analysis.

Description

Technical Field

[0001] The present invention relates to the technical field of AMC over-standard cause analysis, and in particular to an AMC over-standard cause mining and analysis method and system based on an active learning framework. Background Art

[0002] AMCs are airborne molecular contaminants, including organic and inorganic pollutants in cleanroom air, present as gases, vapors, and airborne particles. These contaminants include acids, bases, polymer additives, and organometallic compounds. In microelectronics manufacturing, AMCs can compromise production processes, reduce yields, and impact product quality, necessitating manual analysis.

[0003] Since there are many devices in the AMC generation scenario, the amount of data that needs to be analyzed is very large, and the workload of the manual analysis process is extremely high. Therefore, many automated data analysis solutions have emerged in the prior art. However, these automated data analysis solutions mostly perform comprehensive analysis on the data, and there is not much truly useful data, and the efficiency is still not very high. Therefore, how to provide an AMC exceedance cause analysis solution with a data screening process to improve the efficiency of the exceedance cause analysis is the technical problem that the technical solution of the present invention aims to solve. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and system for mining and analyzing the causes of AMC exceeding the standard based on an active learning framework, so as to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A method and system for mining and analyzing the causes of AMC exceeding the standard based on an active learning framework, the method comprising:

[0007] Acquire monitoring points in the AMC generation scenario, create a statistical matrix based on the locations of the monitoring points, and acquire data at the monitoring points at various moments based on the statistical matrix to obtain a data matrix containing time tags; the monitoring points include at least locations where sensors and signal transceivers are installed;

[0008] Receive the analysis span input by the staff, obtain the data matrix sequence within the analysis span, and calculate the data information amount of any row and column position of the data matrix in the data matrix sequence; the information amount is determined by information entropy and periodicity; the analysis span is a time span;

[0009] The AMC concentrations at the time points obtained by the AMC monitor within the query analysis span are analyzed. When any AMC concentration reaches a preset threshold, a submatrix sequence corresponding to the AMC monitor is extracted from the data matrix sequence. The number of submatrices in the submatrix sequence is a preset value.

[0010] Marking row and column positions in the submatrix sequence where the amount of data information is less than a preset information amount threshold, and deleting the data at the marked row and column positions in the submatrix sequence to obtain a simplified matrix sequence;

[0011] The obtained simplified matrix sequence was used as a feature and the AMC concentration as a label to construct a sample set, and the cause of the exceeding of the standard was determined based on the sample set.

[0012] As a further solution of the present invention, the steps of receiving the analysis span input by the staff, obtaining the data matrix sequence within the analysis span, and calculating the amount of data information at any row and column position of the data matrix in the data matrix sequence include:

[0013] Receive the analysis span input by the staff and obtain the data matrix sequence within the analysis span;

[0014] For any row and column position, read the data at the row and column position in the data matrix sequence in turn, arrange the data based on the sequence order, and obtain the array of each row and column position;

[0015] Perform dimensionless processing on the data in the array and calculate the information entropy;

[0016] Perform periodicity identification on the dimensionless array and determine the probability of the existence of a period;

[0017] The amount of data information is determined based on information entropy and the probability of existence of cycles.

[0018] As a further solution of the present invention, the process of identifying periodicity of the dimensionless array and determining the probability of the existence of a period includes:

[0019] Perform discrete Fourier transform on the dimensionless array to obtain the frequency domain group;

[0020] Calculate the standard deviation of the amplitude of each frequency component in the frequency domain group;

[0021] Fit the amplitude of each frequency component in the frequency domain group into a curve and calculate the curvature of each point on the curve;

[0022] When the curvature reaches a preset threshold, the corresponding frequency is marked;

[0023] When the interval length of the marked frequency is less than a preset length threshold and the standard deviation is greater than a preset standard deviation threshold, the probability of the existence of a cycle is determined based on the marked interval length and the standard deviation.

[0024] As a further solution of the present invention, the query analysis span includes the steps of extracting a submatrix sequence corresponding to the AMC monitor from the data matrix sequence when any AMC concentration reaches a preset threshold value. The steps include:

[0025] Query and analyze the AMC concentration obtained in real time by the AMC monitor within the span; the AMC concentration contains a time tag;

[0026] Comparing the AMC concentration with a preset concentration threshold;

[0027] When any AMC concentration reaches a preset threshold, query the location of the AMC monitor, query the monitoring point closest to the AMC monitor, and query the row and column position corresponding to the monitoring point;

[0028] Determine the extraction side length according to the AMC concentration, expand the queried row and column positions based on the extraction side length, and obtain the position and size of the submatrix; the extraction side length is proportional to the AMC concentration;

[0029] The number of submatrices is determined based on the preset backtracking time, the number of submatrices is selected from the data matrix sequence based on the current moment, the submatrices are intercepted from the data matrix according to the position and size of the submatrices, and the submatrices are arranged in sequence to obtain a submatrix sequence.

[0030] As a further solution of the present invention, the step of marking row and column positions in the submatrix sequence where the amount of data information is less than a preset information amount threshold, and deleting the data at the marked row and column positions in the submatrix sequence to obtain a simplified matrix sequence includes:

[0031] Query the amount of data information at each row and column position in the submatrix sequence;

[0032] Comparing the amount of data information with a preset information amount threshold, and marking the row and column position when the amount of data information is less than the preset information amount threshold;

[0033] For each submatrix in the submatrix sequence, delete the data at the marked row and column positions and regularize the remaining data;

[0034] When each submatrix is regularized, a simplified matrix sequence is obtained.

[0035] As a further solution of the present invention, the steps of using the obtained simplified matrix sequence as a feature and the AMC concentration as a label to construct a sample set and determining the cause of exceeding the standard based on the sample set include:

[0036] The obtained simplified matrix sequence is used as a feature and the AMC concentration is used as a label to construct a sample set;

[0037] Training an AMC concentration prediction model based on the sample set;

[0038] Read the parameters corresponding to each variable in the AMC concentration prediction model and calculate the parameter weights;

[0039] Select variables according to the weight of the parameters, and query the type of monitoring point corresponding to the variable as the reason for exceeding the standard.

[0040] The technical solution of the present invention also provides an AMC excess cause mining and analysis system based on an active learning framework, the system comprising:

[0041] A data statistics module is used to obtain monitoring points in the AMC generation scenario, create a statistical matrix based on the locations of the monitoring points, and obtain data at the monitoring points at various times based on the statistical matrix to obtain a data matrix containing time tags; the monitoring points include at least the locations where sensors and signal transceivers are installed;

[0042] An information quantity calculation module is used to receive the analysis span input by the staff, obtain the data matrix sequence within the analysis span, and calculate the data information quantity of any row and column position of the data matrix in the data matrix sequence; the information quantity is determined by information entropy and periodicity; the analysis span is a time span;

[0043] A submatrix extraction module is used to query and analyze the AMC concentrations at the time points obtained by the AMC monitor within the span. When any AMC concentration reaches a preset threshold, a submatrix sequence corresponding to the AMC monitor is extracted from the data matrix sequence. The number of submatrices in the submatrix sequence is a preset value.

[0044] A matrix simplification module is used to mark the row and column positions in the submatrix sequence where the amount of data information is less than a preset information amount threshold, and delete the data at the marked row and column positions in the submatrix sequence to obtain a simplified matrix sequence;

[0045] The module for analyzing the causes of exceeding the standard is used to construct a sample set using the obtained simplified matrix sequence as a feature and the AMC concentration as a label, and to determine the causes of exceeding the standard based on the sample set.

[0046] As a further solution of the present invention: the information amount calculation module includes:

[0047] A matrix sequence acquisition unit is used to receive the analysis span input by the staff and obtain the data matrix sequence within the analysis span;

[0048] An array extraction unit is used to read the data at any row and column position in the data matrix sequence in sequence, arrange the data based on the sequence order, and obtain an array for each row and column position;

[0049] An information entropy calculation unit is used to perform dimensionless processing on the data in the array and calculate the information entropy;

[0050] The period identification unit is used to identify the periodicity of the dimensionless array and determine the probability of the existence of a period;

[0051] The calculation execution unit is used to determine the amount of data information based on information entropy and the probability of existence cycle.

[0052] As a further solution of the present invention: the submatrix extraction module includes:

[0053] A concentration acquisition unit, used to query and analyze the AMC concentration acquired in real time by the AMC monitor within the span; the AMC concentration contains a time tag;

[0054] A concentration comparison unit, used to compare the AMC concentration with a preset concentration threshold;

[0055] A position query unit is used to query the position of the AMC monitor, the monitoring point closest to the AMC monitor, and the row and column position corresponding to the monitoring point when any AMC concentration reaches a preset threshold;

[0056] a position expansion unit, configured to determine an extraction side length according to the AMC concentration, and expand the queried row and column positions based on the extraction side length to obtain the position and size of the submatrix; wherein the extraction side length is proportional to the AMC concentration;

[0057] The matrix interception unit is used to determine the number of sub-matrices based on a preset backtracking time, select the number of data matrices of sub-matrices in the data matrix sequence based on the current moment, intercept the sub-matrices in the data matrix according to the position and size of the sub-matrices, and arrange the sub-matrices based on the sequence order to obtain a sub-matrix sequence.

[0058] As a further solution of the present invention: the matrix simplification module includes:

[0059] An information amount query unit, used to query the data information amount of each row and column position in the sub-matrix sequence;

[0060] a position marking unit, configured to compare the amount of data information with a preset information amount threshold, and mark the row and column position when the amount of data information is less than the preset information amount threshold;

[0061] A data regularization unit is used to delete the data at the marked row and column positions of each submatrix in the submatrix sequence and regularize the remaining data;

[0062] The sequence output unit is used to obtain a simplified matrix sequence after each sub-matrix is regularized.

[0063] Compared with the prior art, the present invention has the following beneficial effects:

[0064] The present invention provides a data volume-based data screening solution, which actively screens the sample set, reduces the number of samples in the sample set, and improves the quality of samples in the sample set. This preliminary work belongs to the active learning framework and is connected with the existing recognition process, greatly improving the efficiency of cause analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention.

[0066] Figure 1 The overall flow chart of the AMC exceeding standard cause mining and analysis method based on the active learning framework is shown.

[0067] Figure 2 The structural diagram of the AMC excess cause mining and analysis system based on the active learning framework is shown. DETAILED DESCRIPTION

[0068] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0069] Figure 1 The figure is a general flow chart of a method and system for mining and analyzing the causes of AMC exceeding the standard based on an active learning framework. In an embodiment of the present invention, a method for mining and analyzing the causes of AMC exceeding the standard based on an active learning framework is provided. The method includes:

[0070] Step S100: Acquire monitoring points in the AMC generation scenario, create a statistical matrix based on the locations of the monitoring points, and acquire data at the monitoring points at various moments based on the statistical matrix to obtain a data matrix containing time tags; the monitoring points include at least locations where sensors and signal transceivers are installed;

[0071] AMC refers to airborne molecular contaminants, including organic and inorganic pollutants in clean room air, which exist in the form of gas, vapor and airborne dust, such as acids, alkalis, polymer additives, organometallic compounds, etc. In microelectronics manufacturing, AMC can compromise the production process, reduce yield, and affect product quality. AMC scenarios typically involve multiple devices equipped with various sensors for acquiring device data. Furthermore, the scenario also includes environmental monitoring devices, such as temperature measuring instruments. Each device with information collection capabilities corresponds to the monitoring points mentioned above. The positions of all monitoring points are obtained, and a statistical matrix is created based on their positional relationships. The rows and columns in the statistical matrix correspond one-to-one with the monitoring points, and this correspondence remains unchanged. The devices at these monitoring points acquire data in real time, obtaining time-stamped data that is then inserted into the corresponding rows and columns of the statistical matrix, resulting in a data matrix with time stamps. It is worth noting that since the devices corresponding to the monitoring points may be installed at different heights within the scenario, the statistical matrix is actually a three-dimensional matrix. Furthermore, the number of monitoring points in each direction is not necessarily uniform, so there may be some empty positions in the statistical matrix. Default data can be directly inserted into these empty positions.

[0072] Step S200: receiving an analysis span input by a staff member, obtaining a data matrix sequence within the analysis span, and calculating the amount of data information at any row and column position of the data matrix in the data matrix sequence; the amount of information is determined by information entropy and periodicity; the analysis span is a time span;

[0073] The analysis span is a time span input by the staff, such as one day, three days, or one week. All data matrices within the analysis span are queried based on the time tag and sorted in chronological order to obtain a data matrix sequence (a set of data matrices with a sequence). In the data matrix sequence, each data matrix is generated based on a statistical matrix and has the same size. For any row and column position, the amount of data information is calculated based on the data at that row and column position. The independent variables of the calculation process use two parameters: information entropy and periodicity.

[0074] Step S300: querying and analyzing the AMC concentrations at the time points obtained by the AMC monitor within the span. When any AMC concentration reaches a preset threshold, extracting a submatrix sequence corresponding to the AMC monitor from the data matrix sequence; wherein the number of submatrices in the submatrix sequence is a preset value;

[0075] At the same time, there will be multiple AMC monitors in the AMC generation scenario. The monitoring points do not include the locations where the AMC monitors are installed. The monitoring points and AMC monitors are two independent monitoring processes. The AMC concentrations at each moment obtained by the AMC monitor within the query and analysis span are analyzed. When any AMC concentration is large enough, the sub-matrix sequence corresponding to the AMC monitor is extracted from the data matrix sequence. The meaning of this process is that when there is a high concentration of AMC pollutants at a certain location, the monitoring data around it is obtained with the location as the center. Under the premise that the monitoring data of the monitoring points are regarded as the cause, this process is equivalent to simplifying the monitoring data, which is essentially a restriction of the sample and belongs to the active learning framework.

[0076] Step S400: marking the row and column positions in the submatrix sequence where the amount of data information is less than a preset information amount threshold, and deleting the data at the marked row and column positions in the submatrix sequence to obtain a simplified matrix sequence;

[0077] Based on the simplification performed once in step S300, the above content deletes some data again according to the calculated amount of data information, and further simplifies to obtain a simplified matrix sequence.

[0078] Step S500: Using the obtained simplified matrix sequence as a feature and the AMC concentration as a label, a sample set is constructed, and the cause of exceeding the standard is determined based on the sample set;

[0079] The obtained submatrix sequence is the simplified monitoring data. It is used as a feature (independent variable) and the corresponding AMC concentration is used as a label (there is only one AMC concentration, the dependent variable). A sample set is constructed and analyzed to determine the cause of the exceeding of the standard. This process can be assisted by a supervised learning model. Since steps S100 to S400 provide a solution for simplifying the sample set, it is actually also an active learning framework.

[0080] Regarding step S200, the steps of receiving the analysis span input by the staff, obtaining the data matrix sequence within the analysis span, and calculating the data information amount of any row and column position of the data matrix in the data matrix sequence include:

[0081] Receive the analysis span input by the staff and obtain the data matrix sequence within the analysis span;

[0082] For any row and column position, read the data at the row and column position in the data matrix sequence in turn, arrange the data based on the sequence order, and obtain the array of each row and column position;

[0083] Perform dimensionless processing on the data in the array and calculate the information entropy;

[0084] Perform periodicity identification on the dimensionless array and determine the probability of the existence of a period;

[0085] The amount of data information is determined based on information entropy and the probability of existence of cycles.

[0086] Receive the analysis span input by the staff, and obtain the data matrix sequence within the analysis span. For any row and column position, each data matrix in the data matrix sequence has a data at that row and column position. Read these data, arrange them in sequence order, and obtain an array. Since each row and column position corresponds to a monitoring device, there are many types of them, which are troublesome to process together. Therefore, it is necessary to perform a dimensionless processing first. The dimensionless processing method is not complicated. Get the maximum value, calculate the difference between the maximum value and the minimum value, and use it as the denominator. For any data, calculate the difference between it and the minimum value, and use it as the numerator. The ratio obtained is the data after dimensionless processing. Of course, a product constant can also be introduced to adjust the value range of the data after dimensionless processing.

[0087] On this basis, the periodicity of the dimensionless array is identified to determine the probability of the existence of a period, and the amount of data information is determined based on the information entropy and the probability of the existence of a period.

[0088] Furthermore, the calculation process of the information entropy of the array is not complicated and belongs to an existing parameter. Its calculation method is public and will not be described in detail in the present invention. However, the periodic identification process needs to be explained as follows:

[0089] The process of identifying periodicity of the dimensionless array and determining the probability of the existence of a period includes:

[0090] Perform discrete Fourier transform on the dimensionless array to obtain the frequency domain group;

[0091] Calculate the standard deviation of the amplitude of each frequency component in the frequency domain group;

[0092] Fit the amplitude of each frequency component in the frequency domain group into a curve and calculate the curvature of each point on the curve;

[0093] When the curvature reaches a preset threshold, the corresponding frequency is marked;

[0094] When the interval length of the marked frequency is less than a preset length threshold and the standard deviation is greater than a preset standard deviation threshold, the probability of the existence of a cycle is determined based on the marked interval length and the standard deviation.

[0095] By performing a discrete Fourier transform operation, the frequency domain features of the array can be extracted, the standard deviation of the amplitude of each frequency component in the frequency domain group can be calculated, and then the peak data in the frequency domain group can be located. The fewer the peak data (the length of the marked position, corresponding to the frequency interval), the larger the standard deviation, which means that the data in the frequency domain group is more concentrated, the more obvious the periodicity, and the greater the probability of the existence of a period; therefore, the probability of the existence of a period is inversely proportional to the length of the marked interval and directly proportional to the standard deviation. The staff sets the relevant function according to the actual situation, and it only needs to conform to this relationship. The present invention only limits the correlation relationship, not the specific numerical relationship.

[0096] Furthermore, the amount of data information is determined based on information entropy and the probability of the existence of a cycle. The larger the information entropy, the more uncertain the corresponding data is. In the present invention, the more uncertain the data is, the more important it is, that is, the more it needs to be retained as a sample. Therefore, the amount of data information is larger; the greater the probability of the existence of a cycle, the more obvious the periodicity is, the data is considered to be predictable, and its stability is higher. The less important it is, the less it needs to be retained. Therefore, the amount of data information is smaller. In summary, the amount of data information is directly proportional to the information entropy and inversely proportional to the probability of the existence of a cycle. Similar to the above content, the present invention only limits the correlation relationship, not the specific quantitative relationship, because for those skilled in the art of the present invention, many functions can be selected in the existing technology to achieve this relationship.

[0097] Regarding step S300, the query analysis span includes the steps of obtaining the AMC concentrations at the time points obtained by the AMC monitor, and extracting the submatrix sequence corresponding to the AMC monitor from the data matrix sequence when any AMC concentration reaches a preset threshold, including:

[0098] Query and analyze the AMC concentration obtained in real time by the AMC monitor within the span; the AMC concentration contains a time tag;

[0099] Comparing the AMC concentration with a preset concentration threshold;

[0100] When any AMC concentration reaches a preset threshold, query the location of the AMC monitor, query the monitoring point closest to the AMC monitor, and query the row and column position corresponding to the monitoring point;

[0101] Determine the extraction side length according to the AMC concentration, expand the queried row and column positions based on the extraction side length, and obtain the position and size of the submatrix; the extraction side length is proportional to the AMC concentration;

[0102] The number of submatrices is determined based on the preset backtracking time, the number of submatrices is selected from the data matrix sequence based on the current moment, the submatrices are intercepted from the data matrix according to the position and size of the submatrices, and the submatrices are arranged in sequence to obtain a submatrix sequence.

[0103] Query and analyze the AMC concentration obtained by the AMC monitor in real time within the span. When any AMC concentration reaches a sufficiently large value, query the position of the AMC monitor, query the monitoring point closest to the AMC monitor, and query the row and column positions corresponding to the monitoring point. At this time, the corresponding position of the AMC monitor in the data matrix is obtained. The extraction side length is determined according to the AMC concentration. The extraction side length is proportional to the AMC concentration. One calculation method is: ; To extract the side length, is the preset coefficient, For concentration, the calculated extracted side length is an odd number, and the obtained submatrix size is 5*5*5 or 7*7*7. At this time, the position of the submatrix is limited to the center position, and the center position can be the corresponding position of the AMC monitor in the data matrix. Then, starting from the current moment, read the data matrix within the preset backtracking time forward, locate the submatrix in each read data matrix, intercept the data, arrange the submatrix based on the sequence order, and obtain the submatrix sequence.

[0104] The sequence order in the present invention is the order of the data matrix sequence, which is essentially a time order.

[0105] Regarding step S400, the steps of marking row and column positions in the submatrix sequence where the amount of data information is less than a preset information amount threshold, and deleting the data at the marked row and column positions in the submatrix sequence to obtain a simplified matrix sequence include:

[0106] Query the amount of data information at each row and column position in the submatrix sequence;

[0107] Comparing the amount of data information with a preset information amount threshold, and marking the row and column position when the amount of data information is less than the preset information amount threshold;

[0108] For each submatrix in the submatrix sequence, delete the data at the marked row and column positions and regularize the remaining data;

[0109] When each submatrix is regularized, a simplified matrix sequence is obtained.

[0110] The amount of data information at each row and column position in the submatrix sequence is queried. When the amount of data information is small, the data at the corresponding row and column position is considered unimportant. At this time, for each submatrix in the submatrix sequence, the data at the marked row and column positions are deleted. After the data is deleted, some preset default values are inserted at the row and column positions of the deleted data to ensure that the size of the matrix remains unchanged. This belongs to the matrix regularization process. Of course, other data regularization processes can also be provided, which will not be described in detail in the present invention. When each submatrix is regularized, a simplified matrix sequence is obtained.

[0111] As step S500, the steps of constructing a sample set using the obtained simplified matrix sequence as a feature and the AMC concentration as a label, and determining the cause of exceeding the standard based on the sample set include:

[0112] The obtained simplified matrix sequence is used as a feature and the AMC concentration is used as a label to construct a sample set;

[0113] Training an AMC concentration prediction model based on the sample set;

[0114] Read the parameters corresponding to each variable in the AMC concentration prediction model and calculate the parameter weights;

[0115] Select variables according to the weight of the parameters, and query the type of monitoring point corresponding to the variable as the reason for exceeding the standard.

[0116] The obtained simplified matrix sequence is used as a feature, and the AMC concentration is used as a label to construct a sample set. The AMC concentration prediction model is trained based on the sample set. The AMC concentration prediction model is essentially a function. The data in each row and column position is a variable (the variable is in array format, or a vector). Each variable corresponds to a parameter. By comparing the parameters between different variables, the size of the parameter can be determined. The size of the parameter indicates the degree of influence, that is, the weight in the above content; the parameter with a weight greater than the preset threshold is selected, and the type of monitoring point corresponding to the variable is queried as the cause of the exceeding the standard; of course, the variables can also be sorted according to the parameters, and then the types of monitoring points corresponding to the variables are sorted to obtain the importance of the monitoring points, which is equivalent to determining the importance of various causes.

[0117] It is worth mentioning that the AMC concentration prediction model trained in the above content can also be used to predict AMC pollutants. The AMC concentration can be predicted based on the data obtained from the monitoring points. This is an additional function achieved by the technical solution of the present invention.

[0118] Figure 2 The structure diagram of the AMC over-standard cause mining and analysis system based on the active learning framework is shown. In a preferred embodiment of the technical solution of the present invention, a system for mining and analysis of the AMC over-standard cause using the active learning framework is also provided. The system 10 includes:

[0119] The data statistics module 11 is used to obtain monitoring points in the AMC generation scene, create a statistical matrix based on the locations of the monitoring points, and obtain data at the monitoring points at various times based on the statistical matrix to obtain a data matrix containing time tags; the monitoring points include at least the locations where sensors and signal transceivers are installed;

[0120] The information quantity calculation module 12 is used to receive the analysis span input by the staff, obtain the data matrix sequence within the analysis span, and calculate the data information quantity of any row and column position of the data matrix in the data matrix sequence; the information quantity is determined by information entropy and periodicity; the analysis span is a time span;

[0121] The submatrix extraction module 13 is configured to query and analyze the AMC concentrations at the time points obtained by the AMC monitor within the span. When any AMC concentration reaches a preset threshold, the submatrix sequence corresponding to the AMC monitor is extracted from the data matrix sequence. The number of submatrices in the submatrix sequence is a preset value.

[0122] A matrix simplification module 14 is configured to mark row and column positions in the submatrix sequence where the amount of data information is less than a preset information amount threshold, and delete the data at the marked row and column positions in the submatrix sequence to obtain a simplified matrix sequence;

[0123] The exceeding-standard cause analysis module 15 is used to construct a sample set using the obtained simplified matrix sequence as a feature and the AMC concentration as a label, and determine the exceeding-standard cause based on the sample set.

[0124] Furthermore, the information volume calculation module 12 includes:

[0125] A matrix sequence acquisition unit is used to receive the analysis span input by the staff and obtain the data matrix sequence within the analysis span;

[0126] An array extraction unit is used to read the data at any row and column position in the data matrix sequence in sequence, arrange the data based on the sequence order, and obtain an array for each row and column position;

[0127] An information entropy calculation unit is used to perform dimensionless processing on the data in the array and calculate the information entropy;

[0128] The period identification unit is used to identify the periodicity of the dimensionless array and determine the probability of the existence of a period;

[0129] The calculation execution unit is used to determine the amount of data information based on information entropy and the probability of existence cycle.

[0130] Specifically, the submatrix extraction module 13 includes:

[0131] A concentration acquisition unit, used to query and analyze the AMC concentration acquired in real time by the AMC monitor within the span; the AMC concentration contains a time tag;

[0132] A concentration comparison unit, used to compare the AMC concentration with a preset concentration threshold;

[0133] A position query unit is used to query the position of the AMC monitor, the monitoring point closest to the AMC monitor, and the row and column position corresponding to the monitoring point when any AMC concentration reaches a preset threshold;

[0134] a position expansion unit, configured to determine an extraction side length according to the AMC concentration, and expand the queried row and column positions based on the extraction side length to obtain the position and size of the submatrix; wherein the extraction side length is proportional to the AMC concentration;

[0135] The matrix interception unit is used to determine the number of sub-matrices based on a preset backtracking time, select the number of data matrices of sub-matrices in the data matrix sequence based on the current moment, intercept the sub-matrices in the data matrix according to the position and size of the sub-matrices, and arrange the sub-matrices based on the sequence order to obtain a sub-matrix sequence.

[0136] Furthermore, the matrix simplification module 14 includes:

[0137] An information amount query unit, used to query the data information amount of each row and column position in the sub-matrix sequence;

[0138] a position marking unit, configured to compare the amount of data information with a preset information amount threshold, and mark the row and column position when the amount of data information is less than the preset information amount threshold;

[0139] A data regularization unit is used to delete the data at the marked row and column positions of each submatrix in the submatrix sequence and regularize the remaining data;

[0140] The sequence output unit is used to obtain a simplified matrix sequence after each sub-matrix is regularized.

[0141] The above are only preferred embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for mining and analyzing the causes of AMC exceeding the standard based on an active learning framework, characterized in that: The method comprises: Acquire monitoring points in the AMC generation scenario, create a statistical matrix based on the locations of the monitoring points, and acquire data at the monitoring points at various moments based on the statistical matrix to obtain a data matrix containing time tags; the monitoring points include at least locations where sensors and signal transceivers are installed; Receive the analysis span input by the staff, obtain the data matrix sequence within the analysis span, and calculate the data information amount of any row and column position of the data matrix in the data matrix sequence; the information amount is determined by information entropy and periodicity; the analysis span is a time span; Query and analyze the AMC concentration obtained in real time by the AMC monitor within the span; the AMC concentration contains a time tag; Comparing the AMC concentration with a preset concentration threshold; When any AMC concentration reaches a preset threshold, query the location of the AMC monitor, query the monitoring point closest to the AMC monitor, and query the row and column position corresponding to the monitoring point; Determine the extraction side length according to the AMC concentration, expand the queried row and column positions based on the extraction side length, and obtain the position and size of the submatrix; the extraction side length is proportional to the AMC concentration; The number of submatrices is determined based on a preset backtracking time, a data matrix of the number of submatrices is selected from the data matrix sequence based on the current moment, a submatrix is intercepted from the data matrix according to the position and size of the submatrix, and the submatrix is arranged in sequence to obtain a submatrix sequence; Marking row and column positions in the submatrix sequence where the amount of data information is less than a preset information amount threshold, and deleting the data at the marked row and column positions in the submatrix sequence to obtain a simplified matrix sequence; The obtained simplified matrix sequence is used as a feature and the AMC concentration is used as a label to construct a sample set; Training an AMC concentration prediction model based on the sample set; Read the parameters corresponding to each variable in the AMC concentration prediction model and calculate the parameter weights; Select variables based on parameter weights, and query the type of monitoring point corresponding to the variable as the cause of exceeding the standard; Among them, AMC refers to airborne molecular contaminants.

2. The method for mining and analyzing the causes of AMC exceeding the standard based on the active learning framework according to claim 1 is characterized in that: The steps of receiving the analysis span input by the staff, obtaining the data matrix sequence within the analysis span, and calculating the data information amount of any row and column position of the data matrix in the data matrix sequence include: Receive the analysis span input by the staff and obtain the data matrix sequence within the analysis span; For any row and column position, read the data at the row and column position in the data matrix sequence in turn, arrange the data based on the sequence order, and obtain the array of each row and column position; Perform dimensionless processing on the data in the array and calculate the information entropy; Perform periodicity identification on the dimensionless array and determine the probability of the existence of a period; The amount of data information is determined based on information entropy and the probability of existence of cycles.

3. The method for mining and analyzing the causes of AMC exceeding the standard based on the active learning framework according to claim 2 is characterized in that: The process of identifying periodicity of the dimensionless array and determining the probability of the existence of a period includes: Perform discrete Fourier transform on the dimensionless array to obtain the frequency domain group; Calculate the standard deviation of the amplitude of each frequency component in the frequency domain group; Fit the amplitude of each frequency component in the frequency domain group into a curve and calculate the curvature of each point on the curve; When the curvature reaches a preset threshold, the corresponding frequency is marked; When the interval length of the marked frequency is less than a preset length threshold and the standard deviation is greater than a preset standard deviation threshold, the probability of the existence of a cycle is determined based on the marked interval length and the standard deviation.

4. The method for mining and analyzing the causes of AMC exceeding the standard based on the active learning framework according to claim 1 is characterized in that: The step of marking row and column positions in the submatrix sequence where the amount of data information is less than a preset information amount threshold, and deleting the data at the marked row and column positions in the submatrix sequence to obtain a simplified matrix sequence includes: Query the amount of data information at each row and column position in the submatrix sequence; Comparing the amount of data information with a preset information amount threshold, and marking the row and column position when the amount of data information is less than the preset information amount threshold; For each submatrix in the submatrix sequence, delete the data at the marked row and column positions and regularize the remaining data; When each submatrix is regularized, a simplified matrix sequence is obtained.

5. An active learning framework AMC over-standard cause mining and analysis system, characterized by: The system comprises: A data statistics module is used to obtain monitoring points in the AMC generation scenario, create a statistical matrix based on the locations of the monitoring points, and obtain data at the monitoring points at various times based on the statistical matrix to obtain a data matrix containing time tags; the monitoring points include at least the locations where sensors and signal transceivers are installed; An information quantity calculation module is used to receive the analysis span input by the staff, obtain the data matrix sequence within the analysis span, and calculate the data information quantity of any row and column position of the data matrix in the data matrix sequence; the information quantity is determined by information entropy and periodicity; the analysis span is a time span; A concentration acquisition unit, used to query and analyze the AMC concentration acquired in real time by the AMC monitor within the span; the AMC concentration contains a time tag; A concentration comparison unit, used to compare the AMC concentration with a preset concentration threshold; A position query unit is used to query the position of the AMC monitor, the monitoring point closest to the AMC monitor, and the row and column position corresponding to the monitoring point when any AMC concentration reaches a preset threshold; a position expansion unit, configured to determine an extraction side length according to the AMC concentration, and expand the queried row and column positions based on the extraction side length to obtain the position and size of the submatrix; wherein the extraction side length is proportional to the AMC concentration; a matrix interception unit, configured to determine the number of submatrices based on a preset backtracking time, select a data matrix having the same number of submatrices as the submatrices in the data matrix sequence based on the current moment, intercept the submatrices in the data matrix according to the positions and sizes of the submatrices, and arrange the submatrices in a sequence order to obtain a submatrix sequence; A matrix simplification module is used to mark the row and column positions in the submatrix sequence where the amount of data information is less than a preset information amount threshold, and delete the data at the marked row and column positions in the submatrix sequence to obtain a simplified matrix sequence; The module for analyzing the causes of exceeding the standard is used to construct a sample set using the obtained simplified matrix sequence as a feature and the AMC concentration as a label, and to determine the causes of exceeding the standard based on the sample set; Among them, AMC refers to airborne molecular contaminants; The simplified matrix sequence obtained is used as a feature, the AMC concentration is used as a label, a sample set is constructed, and the content of determining the cause of exceeding the standard based on the sample set includes: The obtained simplified matrix sequence is used as a feature and the AMC concentration is used as a label to construct a sample set; Training an AMC concentration prediction model based on the sample set; Read the parameters corresponding to each variable in the AMC concentration prediction model and calculate the parameter weights; Select variables according to the weight of the parameters, and query the type of monitoring point corresponding to the variable as the reason for exceeding the standard.

6. The AMC exceeding standard cause mining and analysis system based on the active learning framework according to claim 5 is characterized in that: The information volume calculation module includes: A matrix sequence acquisition unit is used to receive the analysis span input by the staff and obtain the data matrix sequence within the analysis span; An array extraction unit is used to read the data at any row and column position in the data matrix sequence in sequence, arrange the data based on the sequence order, and obtain an array for each row and column position; An information entropy calculation unit is used to perform dimensionless processing on the data in the array and calculate the information entropy; The period identification unit is used to identify the periodicity of the dimensionless array and determine the probability of the existence of a period; The calculation execution unit is used to determine the amount of data information based on information entropy and the probability of existence cycle.

7. The AMC exceeding standard cause mining and analysis system based on the active learning framework according to claim 5 is characterized in that: The matrix simplification module includes: An information amount query unit, used to query the data information amount of each row and column position in the sub-matrix sequence; a position marking unit, configured to compare the amount of data information with a preset information amount threshold, and mark the row and column position when the amount of data information is less than the preset information amount threshold; A data regularization unit is used to delete the data at the marked row and column positions of each submatrix in the submatrix sequence and regularize the remaining data; The sequence output unit is used to obtain a simplified matrix sequence after each sub-matrix is regularized.

Images