CN112699547A - Sliding window type multivariate time sequence missing value filling method based on 5G network - Google Patents
Sliding window type multivariate time sequence missing value filling method based on 5G network Download PDFInfo
- Publication number
- CN112699547A CN112699547A CN202011565488.4A CN202011565488A CN112699547A CN 112699547 A CN112699547 A CN 112699547A CN 202011565488 A CN202011565488 A CN 202011565488A CN 112699547 A CN112699547 A CN 112699547A
- Authority
- CN
- China
- Prior art keywords
- data
- window
- filling
- value
- time
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/30—Services specially adapted for particular environments, situations or purposes
- H04W4/38—Services specially adapted for particular environments, situations or purposes for collecting sensor information
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/12—Timing analysis or timing optimisation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Pure & Applied Mathematics (AREA)
- Computational Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Databases & Information Systems (AREA)
- Computing Systems (AREA)
- Algebra (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Testing Or Calibration Of Command Recording Devices (AREA)
Abstract
The invention provides a sliding window type multivariate time sequence missing value filling method and device based on a 5G network, wherein the sliding window type multivariate time sequence missing value filling method based on the 5G network collects sensor data of each measuring point of equipment; merging the sensor data of each measuring point; filling missing values in a matrix iterative optimization mode on the basis of reconstruction errors for time series data fragments with fixed window sizes according to preset window values; and filling missing values in the whole time sequence data by using a sliding window according to a preset step value. The method solves the problems of low network bandwidth and long time delay, makes full use of the relation between data of various measuring points in historical operation of equipment by using an alternating matrix algorithm for proportion and a sliding window method, and has higher accuracy in filling missing values of data compared with mean filling, maximum filling, mode filling and model prediction filling. Meanwhile, the method provided by the invention is simple to implement, has high portability and is suitable for most of equipment.
Description
Technical Field
The invention belongs to the technical field of data filling, and particularly relates to a sliding window type multi-element time sequence missing value filling method and device based on a 5G network.
Background
With the rapid development of the industrial internet of things technology, more and more devices are connected to the internet of things platform through various sensors. Due to sensor communication abnormity caused by sensor damage, low traditional network bandwidth, time delay and the like, the problems of data loss and loss easily occur in the acquisition process of industrial data. And the existence of the missing value can bring difficulty to the subsequent analysis of industrial data. Therefore, the method has very important significance for solving the problem of industrial data set loss caused by the conditions of low bandwidth, time delay and the like of the traditional network.
The currently common solution is to increase the traditional network bandwidth and adopt the traditional filling method including: mean filling, pre/post value filling, nearest neighbor filling, random forest filling, and the like. But this solution increases the industrial data acquisition cost while ignoring the correlation between different time series data when data population is performed.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: a sliding window type multi-element time sequence missing value filling method and device based on a 5G network are provided.
The technical scheme adopted by the invention for solving the technical problems is as follows: a sliding window type multivariate time sequence missing value filling method based on a 5G network is provided, and the method comprises the following steps:
collecting sensor data of each measuring point of the equipment, and transmitting the sensor data through a 5G network;
merging the sensor data of each measuring point to form multivariate time sequence data;
filling missing values in a matrix iterative optimization mode on the basis of reconstruction errors for time series data fragments with fixed window sizes according to preset window values;
and filling missing values in the whole time sequence data by using a sliding window according to a preset step value.
Further, the method for combining the sensor data of the measuring points to form the multivariate time sequence data comprises the following steps:
aligning the sensor data with a time axis of the time data;
and combining the sensor data of the measuring points according to a time axis to form time sequence data.
Further, the method for aligning the sensor data with the time axis of the time data includes: and solving the union set of the timestamps and arranging the timestamps according to time to obtain a final time axis.
Further, the method for filling missing values in a matrix iterative optimization manner based on reconstruction errors for time series data segments with fixed window sizes according to a preset step value and a preset window value includes:
setting the size of a window, and intercepting a time sequence segment;
for the time sequence data segment with the fixed window size, the window size is n, the number of the measuring points is m, and the multivariate time sequence segment can be expressed as a matrix A belonging to Rm×nThe multiple time sequence segment has missing values, and the matrix A is a sparse matrix containing the missing values;
for matrix A ∈ Rm×nDecomposable into matrices U and VTProduct of (a ≈ UV)TWherein U ∈ Rm×k，V∈Rn×k，k＜m，n；
Alternately solving matrixes U and V according to the minimized reconstruction error as an optimization target;
after the matrices U and V have been determined, UV is usedTCorresponding element in (1)The element fills in the missing values in matrix a.
Further, the method for alternately solving the matrices U and V according to the minimization of the reconstruction error as the optimization objective includes:
the reconstruction error is:
wherein, aijIs a known element in the ith row and the jth column of the matrix A, uiIs the i-th row vector, v, of the matrix UjIs the jth row vector of the matrix V, and lambda is the regularization coefficient;
and (3) minimizing the reconstruction error to obtain an optimal solution of U and V: u shape*，V*＝argminU，VL(U，V)；
And assigning an initial value to U, solving V by using a least square method, fixing V, solving U by using the same method, and then alternately iterating until a preset convergence condition is met, namely the reconstruction error is minimum.
Further, the method for filling missing values in the entire time series data by using a sliding window according to the preset step value comprises the following steps:
setting the sliding step of the window as S, limiting the sliding step W/2 of the window to be more than S and less than W, and setting W as the boundary value of the window;
carrying out matrix decomposition and reconstruction on the data in the single window;
and filling missing values in the data according to the reconstruction result. An overlapping area with the width of W-S exists between two continuous windows, and if missing values exist in data of the overlapping area, the average value is taken as a final filling value;
and sequentially sliding the window, traversing the whole time sequence data, and completing the filling of missing values in the whole data.
The invention also provides a sliding window type multivariate time sequence missing value filling device based on the 5G network, which comprises:
the data acquisition module is suitable for acquiring sensor data of each measuring point of the equipment and transmitting the sensor data through a 5G network;
the data processing module is suitable for combining the sensor data of each measuring point to form multivariate time sequence data;
the fixed window missing value supplementing module is suitable for filling missing values in a matrix iterative optimization mode on the basis of reconstruction errors for the time series data fragments with the fixed window size according to a preset window value;
and the sliding window missing value supplementing module is suitable for filling missing values in the whole time sequence data by using a sliding window according to a preset step value.
Further, the sensor data of each measuring point acquired by the data acquisition module is unary time series data.
The invention also provides a computer-readable storage medium, wherein one or more instructions are stored in the computer-readable storage medium, and the processor of the apparatus for risk analysis in the one or more instructions executes the sliding-window multivariate time-series missing value filling method based on the 5G network.
The present invention also provides an electronic device, comprising: a memory and a processor; at least one program instruction is stored in the memory; the processor implements the 5G network-based sliding-window multi-element time-series missing value filling method as described above by loading and executing the at least one program instruction.
The invention has the beneficial effects that: the invention provides a sliding window type multivariate time sequence missing value filling method and device based on a 5G network, wherein the sliding window type multivariate time sequence missing value filling method based on the 5G network collects sensor data of each measuring point of equipment and transmits the sensor data through the 5G network; merging the sensor data of each measuring point to form multivariate time sequence data; filling missing values in a matrix iterative optimization mode on the basis of reconstruction errors for time series data fragments with fixed window sizes according to preset window values; and filling missing values in the whole time sequence data by using a sliding window according to a preset step value. The method solves the problems of low network bandwidth and long time delay, makes full use of the relation between data of various measuring points in historical operation of equipment by using an alternating matrix algorithm for proportion and a sliding window method, and has higher accuracy in filling missing values of data compared with mean filling, maximum filling, mode filling and model prediction filling. Meanwhile, the method provided by the invention is simple to implement, has high portability and is suitable for most of equipment.
Drawings
The invention is further illustrated by the following figures and examples.
Fig. 1 is a flowchart of a sliding-window multi-element time-series missing value filling method based on a 5G network according to an embodiment of the present invention.
Fig. 2 is a schematic block diagram of a sliding-window multi-element time-series missing value filling apparatus based on a 5G network according to an embodiment of the present invention.
Fig. 3 is a partial block diagram of an electronic device provided by an embodiment of the invention.
Detailed Description
The present invention will now be described in detail with reference to the accompanying drawings. This figure is a simplified schematic diagram, and merely illustrates the basic structure of the present invention in a schematic manner, and therefore it shows only the constitution related to the present invention.
Example 1
Referring to fig. 1, the present invention provides a sliding window type multi-element time sequence missing value filling method based on a 5G network. The method solves the problems of low network bandwidth and long time delay, makes full use of the relation between data of various measuring points in historical operation of equipment by using an alternating matrix algorithm for proportion and a sliding window method, and has higher accuracy in filling missing values of data compared with mean filling, maximum filling, mode filling and model prediction filling. Meanwhile, the method provided by the invention is simple to implement, has high portability and is suitable for most of equipment.
Specifically, the sliding window type multivariate time sequence missing value filling method based on the 5G network comprises the following steps:
s110: and collecting sensor data of each measuring point of the equipment, and transmitting the sensor data through a 5G network.
Specifically, historical operating parameter data of the device, taking the pressure fan as an example, is collected, historical data of each measuring point of the pressure fan, including 12 measuring points such as current, power, exhaust pressure, exhaust temperature, rotating speed value, and the like, and the final data is a plurality of unary time series data, which is described below with reference to current and rotating speed data, in the displayed time series segment, a missing value exists in each of the current data and the rotating speed data, as shown in table 1 and table 2.
Table 1: wind turbine current data
| Time | 08:00:00 | 08:00:30 | 08:01:00 | 08:02:00 | 08:02:30 | … |
| Electric current | 21.32 | 21.88 | 22.09 | 21.50 | 22.21 | … |
Table 2: blower speed data
| Time | 08:00:00 | 08:01:00 | 08:01:30 | 08:02:00 | 08:02:30 | … |
| Rotational speed | 2807 | 2873 | 2786 | 2751 | 2887 | … |
S120: and combining the sensor data of the measuring points to form a plurality of time sequence data.
Specifically, step S120 includes the following sub-steps:
s121: the sensor data is aligned with the time axis of the time data.
Specifically, due to sensor malfunction or other reasons, the data acquisition and storage process may cause data missing, and when a plurality of time-series data containing missing values are combined into a plurality of time-series data, their time axes are aligned first. The alignment time axis is a union set of the time data of each time series data and is arranged in sequence, and the results of the alignment operation performed on the current and rotation speed data segments shown in table 1 and table 2 are shown in table 3.
Table 3: forced air machine current and speed data set
| Time | 08:00:00 | 08:00:30 | 08:01:00 | 08:01:30 | 08:02:00 | 08:02:30 | … |
| Electric current | 21.32 | 21.88 | 22.09 | 21.50 | 22.21 | … | |
| Rotational speed | 2807 | 2883 | 2786 | 2826 | 2790 | … |
S122: and combining the sensor data of the measuring points according to a time axis to form time sequence data.
S130: and filling missing values in a matrix iterative optimization mode on the basis of reconstruction errors for the time series data fragments with the fixed window size according to a preset window value.
Specifically, step S130 includes the steps of:
s131: and setting the window size and intercepting the time sequence fragments.
Specifically, the window size is set appropriately according to the details of the data set, such as the length of time-series data and the number of stations. In this example, the window width W is set to 60
S132: for the time sequence data segment with the fixed window size, the window size is n, the number of the measuring points is m, and the multivariate time sequence segment can be expressed as a matrix A belonging to Rm×nThe multiple time sequence segment has missing values, and the matrix A is a sparse matrix containing the missing values; in this example, n ═ W ═ 60, and m ═ 12, so a ∈ R12×60As shown in table 4.
Table 4: multivariate time series data fragment with window size of 60
S133: for matrix A ∈ Rm×nDecomposable into matrices U and VTProduct of (a ≈ UV)TWherein U ∈ Rm×k，V∈Rn ×k，k＜m，n。
In particular toIn other words, the matrix A ∈ R in Table 4 is compared12×60Can be approximately decomposed into matrices U and VTIf k is 5, then U is equal to R12×5，V∈R60×5。
S134: and alternately solving the matrixes U and V according to the minimized reconstruction error as an optimization target.
Specifically, step S134 includes the steps of:
s1341: the reconstruction error is:
wherein, aijIs a known element in the ith row and the jth column of the matrix A, uiIs the i-th row vector, v, of the matrix UjIs the jth row vector of the matrix V, and lambda is the regularization coefficient;
s1342: and (3) minimizing the reconstruction error to obtain an optimal solution of U and V: u shape*，V*＝arg minU，VL(U，V)；
S1343: and assigning an initial value to U, solving V by using a least square method, fixing V, solving U by using the same method, and then alternately iterating until a preset convergence condition is met, namely the reconstruction error is minimum.
S135: after the matrices U and V have been determined, UV is usedTThe corresponding element in (a) fills the missing value in matrix a.
In this embodiment, UV can be used after the matrices U and V are obtained from the data of Table 4TThe corresponding elements in (a) fill in the missing values in the matrix a, and the final filling result is shown in table 5.
Table 5: populated multivariate time series data fragments
S140: and filling missing values in the whole time sequence data by using a sliding window according to a preset step value.
Specifically, step S140 includes the steps of:
s141: setting the sliding step of the window as S, limiting the sliding step W/2 of the window to be more than S and less than W, wherein W is the boundary value of the window.
In this embodiment, if the step S is 30, an overlapped region with a width of 30 exists in two consecutive time windows, the missing value in the overlapped region is calculated twice, and the average value of the two values is used as the final filling value.
S142: carrying out matrix decomposition and reconstruction on the data in the single window;
s143: and filling missing values in the data according to the reconstruction result. An overlapping area with the width of W-S exists between two continuous windows, and if missing values exist in data of the overlapping area, the average value is taken as a final filling value;
s144: and sequentially sliding the window, traversing the whole time sequence data, and completing the filling of missing values in the whole data.
In this embodiment, the missing value padding effect is evaluated using Normalized Root Mean Square Error (NRMSE), which is defined as:
wherein, ymaxAnd yminThe maximum and minimum values of the true values, respectively.
In this embodiment, missing values in proportions of 0.1, 0.2, 0.3, 0.4 and 0.5 are respectively introduced into the pressure fan data set, the window size is set to be 60, the step length is 30, k is 5, data under different missing proportions are filled, the filling effect is evaluated by using NRMSE, and the result is compared with other missing value filling methods, as shown in table 6, the smaller the value of NRMSE represents the better the filling effect.
Table 6: evaluation results of different methods (smaller represents better effect)
| The invention | Mean value filling | Nearest neighbor filling | Random forest filling | Proportion of deficiency |
| 0.74117 | 1.0349 | 0.89573 | 0.76960 | 0.1 |
| 0.75065 | 1.0252 | 0.91066 | 0.77365 | 0.2 |
| 0.75739 | 1.0259 | 0.93899 | 0.78276 | 0.3 |
| 0.77898 | 1.0160 | 0.98302 | 0.79619 | 0.4 |
| 0.78357 | 1.0067 | 1.01978 | 0.80658 | 0.5 |
As can be seen from Table 6, the method of the present invention has the best filling effect at different ratios of the missing values.
Example 2
Referring to fig. 2, the present embodiment provides a sliding-window-type multivariate time-series missing value filling apparatus based on a 5G network, the apparatus includes:
the data acquisition module is suitable for acquiring sensor data of each measuring point of the equipment and transmitting the sensor data through a 5G network;
and the data processing module is suitable for combining the sensor data of each measuring point to form multi-element time sequence data.
The sensor data of each measuring point acquired by the data acquisition module is unary time series data.
The fixed window missing value supplementing module is suitable for filling missing values in a matrix iterative optimization mode on the basis of reconstruction errors for the time series data fragments with the fixed window size according to a preset window value;
and the sliding window missing value supplementing module is suitable for filling missing values in the whole time sequence data by using a sliding window according to a preset step value.
Example 3
Embodiments of the present invention also provide a computer-readable storage medium having one or more instructions stored therein, where a processor of a device for risk analysis within the one or more instructions executes a sliding-window multivariate time-series missing value filling method based on a 5G network as provided in embodiment 1.
In the embodiment, the sliding window type multivariate time sequence missing value filling method based on the 5G network collects the sensor data of each measuring point of the equipment and transmits the sensor data through the 5G network; merging the sensor data of each measuring point to form multivariate time sequence data; filling missing values in a matrix iterative optimization mode on the basis of reconstruction errors for time series data fragments with fixed window sizes according to preset window values; and filling missing values in the whole time sequence data by using a sliding window according to a preset step value. The method solves the problems of low network bandwidth and long time delay, makes full use of the relation between data of various measuring points in historical operation of equipment by using an alternating matrix algorithm for proportion and a sliding window method, and has higher accuracy in filling missing values of data compared with mean filling, maximum filling, mode filling and model prediction filling. Meanwhile, the method provided by the invention is simple to implement, has high portability and is suitable for most of equipment.
Example 4
Referring to fig. 3, an embodiment of the present invention further provides an electronic device, including: a memory and a processor; at least one program instruction is stored in the memory; the processor loads and executes the at least one program instruction to realize the 5G network-based sliding window type multivariate time sequence missing value filling method provided by the embodiment 1.
The memory 502 and the processor 501 are coupled in a bus that may include any number of interconnected buses and bridges that couple one or more of the various circuits of the processor 501 and the memory 502 together. The bus may also connect various other circuits such as peripherals, voltage regulators, power management circuits, and the like, which are well known in the art, and therefore, will not be described any further herein. A bus interface provides an interface between the bus and the transceiver. The transceiver may be one element or a plurality of elements, such as a plurality of receivers and transmitters, providing a means for communicating with various other apparatus over a transmission medium. The data processed by the processor 501 is transmitted over a wireless medium through an antenna, which further receives the data and transmits the data to the processor 501.
The processor 501 is responsible for managing the bus and general processing and may also provide various functions including timing, peripheral interfaces, voltage regulation, power management, and other control functions. And memory 502 may be used to store data used by processor 501 in performing operations.
In summary, the invention provides a sliding window type multivariate time sequence missing value filling method and device based on a 5G network, wherein the sliding window type multivariate time sequence missing value filling method based on the 5G network collects sensor data of each measuring point of equipment and transmits the sensor data through the 5G network; merging the sensor data of each measuring point to form multivariate time sequence data; filling missing values in a matrix iterative optimization mode on the basis of reconstruction errors for time series data fragments with fixed window sizes according to preset window values; and filling missing values in the whole time sequence data by using a sliding window according to a preset step value. The method solves the problems of low network bandwidth and long time delay, makes full use of the relation between data of various measuring points in historical operation of equipment by using an alternating matrix algorithm for proportion and a sliding window method, and has higher accuracy in filling missing values of data compared with mean filling, maximum filling, mode filling and model prediction filling. Meanwhile, the method provided by the invention is simple to implement, has high portability and is suitable for most of equipment.
In light of the foregoing description of preferred embodiments in accordance with the invention, it is to be understood that numerous changes and modifications may be made by those skilled in the art without departing from the scope of the invention. The technical scope of the present invention is not limited to the contents of the specification, and must be determined according to the scope of the claims.
Claims (10)
1. A sliding window type multivariate time sequence missing value filling method based on a 5G network is characterized by comprising the following steps:
collecting sensor data of each measuring point of the equipment, and transmitting the sensor data through a 5G network;
merging the sensor data of each measuring point to form multivariate time sequence data;
filling missing values in a matrix iterative optimization mode on the basis of reconstruction errors for time series data fragments with fixed window sizes according to preset window values;
and filling missing values in the whole time sequence data by using a sliding window according to a preset step value.
2. The sliding-window multivariate time-series missing value filling method based on the 5G network as claimed in claim 1, wherein the method for combining the sensor data of each measuring point to form multivariate time-series data comprises the following steps:
aligning the sensor data with a time axis of the time data;
and combining the sensor data of the measuring points according to a time axis to form time sequence data.
3. The 5G network-based sliding-window multivariate time-series missing value filling method of claim 2, wherein the method of aligning the sensor data with the time axis of the time data is: and solving the union set of the timestamps and arranging the timestamps according to time to obtain a final time axis.
4. The sliding-window multi-element time-series missing value filling method based on the 5G network as claimed in claim 1, wherein the method for filling missing values based on reconstruction errors using matrix iterative optimization for time-series data segments with fixed window size according to the preset window value according to the preset step value comprises:
setting the size of a window, and intercepting a time sequence segment;
for the time sequence data segment with the fixed window size, the window size is n, the number of the measuring points is m, and the multivariate time sequence segment can be expressed as a matrix A belonging to Rm×nThe multiple time sequence segment has missing values, and the matrix A is a sparse matrix containing the missing values;
for matrix A ∈ Rm×nDecomposable into matrices U and VTProduct of (a ≈ UV)TWherein U ∈ Rm×k，V∈Rn×k，k＜m，n；
Alternately solving matrixes U and V according to the minimized reconstruction error as an optimization target;
after the matrices U and V have been determined, UV is usedTThe corresponding element in (a) fills the missing value in matrix a.
5. The sliding-window multivariate time-series missing value filling method based on the 5G network as claimed in claim 4, wherein the alternating matrix U and V solving method based on minimizing reconstruction error as an optimization goal comprises:
the reconstruction error is:
wherein, aijIs a known element in the ith row and the jth column of the matrix A, uiIs the i-th row vector, v, of the matrix UjIs the jth row vector of the matrix V, and lambda is the regularization coefficient;
and (3) minimizing the reconstruction error to obtain an optimal solution of U and V: u shape*，V*＝arg minU，VL(U，V)；
And assigning an initial value to U, solving V by using a least square method, fixing V, solving U by using the same method, and then alternately iterating until a preset convergence condition is met, namely the reconstruction error is minimum.
6. The sliding-window-type multivariate time-series missing-value filling method based on the 5G network as claimed in claim 4, wherein the method for filling missing values in the whole time-series data by using the sliding window according to the preset step value comprises the following steps:
setting the sliding step of the window as S, limiting the sliding step W/2 of the window to be more than S and less than W, and setting W as the boundary value of the window;
carrying out matrix decomposition and reconstruction on the data in the single window;
and filling missing values in the data according to the reconstruction result. An overlapping area with the width of W-S exists between two continuous windows, and if missing values exist in data of the overlapping area, the average value is taken as a final filling value;
and sequentially sliding the window, traversing the whole time sequence data, and completing the filling of missing values in the whole data.
7. A sliding-window multivariate time-series missing value filling apparatus based on a 5G network, the apparatus comprising:
the data acquisition module is suitable for acquiring sensor data of each measuring point of the equipment and transmitting the sensor data through a 5G network;
the data processing module is suitable for combining the sensor data of each measuring point to form multivariate time sequence data;
the fixed window missing value supplementing module is suitable for filling missing values in a matrix iterative optimization mode on the basis of reconstruction errors for the time series data fragments with the fixed window size according to a preset window value;
and the sliding window missing value supplementing module is suitable for filling missing values in the whole time sequence data by using a sliding window according to a preset step value.
8. The sliding-window multivariate time-series missing value filling device based on the 5G network as claimed in claim 7, wherein the sensor data of each measuring point collected by the data collection module is univariate time-series data.
9. A computer readable storage medium having one or more instructions stored therein, wherein a processor of an apparatus for risk analysis within the one or more instructions, when executed, implements the sliding-window multivariate time-series missing value population method based on a 5G network of any one of claims 1-7.
10. An electronic device, comprising: a memory and a processor; at least one program instruction is stored in the memory; the processor implements the 5G network-based sliding-window multi-element time-series missing value population method of any of claims 1-7 by loading and executing the at least one program instruction.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011565488.4A CN112699547A (en) | 2020-12-25 | 2020-12-25 | Sliding window type multivariate time sequence missing value filling method based on 5G network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011565488.4A CN112699547A (en) | 2020-12-25 | 2020-12-25 | Sliding window type multivariate time sequence missing value filling method based on 5G network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112699547A true CN112699547A (en) | 2021-04-23 |
Family
ID=75510714
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011565488.4A Pending CN112699547A (en) | 2020-12-25 | 2020-12-25 | Sliding window type multivariate time sequence missing value filling method based on 5G network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112699547A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113486074A (en) * | 2021-06-02 | 2021-10-08 | 北方工业大学 | Method for acquiring delay time of related process parameter |
| CN117849302A (en) * | 2024-03-08 | 2024-04-09 | 深圳市朗石科学仪器有限公司 | Multi-parameter water quality on-line monitoring method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110175168A (en) * | 2019-05-28 | 2019-08-27 | 山东大学 | A kind of time series data complementing method and system based on generation confrontation network |
| CN110457867A (en) * | 2019-10-08 | 2019-11-15 | 杭州知衣科技有限公司 | A kind of time series data based on machine learning is filled up and restoring method |
| CN111859275A (en) * | 2020-07-20 | 2020-10-30 | 厦门大学 | Mass spectrum data missing value filling method and system based on non-negative matrix factorization |
| CN111881413A (en) * | 2020-07-28 | 2020-11-03 | 中国人民解放军海军航空大学 | Multi-source time sequence missing data recovery method based on matrix decomposition |
| WO2020250247A2 (en) * | 2019-06-13 | 2020-12-17 | Tata Consultancy Services Limited | Method and system for industrial anomaly detection |
-
2020
- 2020-12-25 CN CN202011565488.4A patent/CN112699547A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110175168A (en) * | 2019-05-28 | 2019-08-27 | 山东大学 | A kind of time series data complementing method and system based on generation confrontation network |
| WO2020250247A2 (en) * | 2019-06-13 | 2020-12-17 | Tata Consultancy Services Limited | Method and system for industrial anomaly detection |
| CN110457867A (en) * | 2019-10-08 | 2019-11-15 | 杭州知衣科技有限公司 | A kind of time series data based on machine learning is filled up and restoring method |
| CN111859275A (en) * | 2020-07-20 | 2020-10-30 | 厦门大学 | Mass spectrum data missing value filling method and system based on non-negative matrix factorization |
| CN111881413A (en) * | 2020-07-28 | 2020-11-03 | 中国人民解放军海军航空大学 | Multi-source time sequence missing data recovery method based on matrix decomposition |
Non-Patent Citations (3)
| Title |
|---|
| 汪灿等: "基于交替最小二乘法的时空张量填充算法", 《电气技术》 * |
| 王新恒等: "RW-MC:基于随机游走的自适应矩阵填充算法", 《通信学报》 * |
| 黄晓宇等: "面向时序数据的矩阵分解?", 《软件学报》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113486074A (en) * | 2021-06-02 | 2021-10-08 | 北方工业大学 | Method for acquiring delay time of related process parameter |
| CN113486074B (en) * | 2021-06-02 | 2023-11-03 | 北方工业大学 | Method for acquiring related process parameter delay time |
| CN117849302A (en) * | 2024-03-08 | 2024-04-09 | 深圳市朗石科学仪器有限公司 | Multi-parameter water quality on-line monitoring method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112699547A (en) | Sliding window type multivariate time sequence missing value filling method based on 5G network | |
| WO2021129086A1 (en) | Traffic prediction method, device, and storage medium | |
| CN110475224A (en) | A kind of sensing data processing and collaborative forecasting method based on edge calculations | |
| US11881013B2 (en) | Video system | |
| CN108435819B (en) | Energy consumption abnormity detection method for aluminum profile extruder | |
| US20220335179A1 (en) | System and Method for Calibrating a Model of Thermal Dynamics | |
| CN112383891A (en) | Equipment registration method and device based on object model automatic matching | |
| CN111814344A (en) | Overhead transmission line current-carrying capacity calculation method, system and equipment | |
| CN106788885B (en) | Electrocardiosignal compression rate self-adaptive adjustment wireless transmission system and transmission method | |
| CN111652375A (en) | Intelligent detection and diagnosis method and device for cooling coil faults based on Bayesian inference and virtual sensing | |
| CN105743611B (en) | A kind of wireless sensor network missing data reconstruction method based on sparse dictionary | |
| CN115186699A (en) | Method, device, equipment, system and medium for managing health of railway passenger station equipment | |
| CN114896216A (en) | Industrial Internet data storage method and system based on block chain and electronic equipment | |
| CN113624998A (en) | Electric boiler heat supplementing and heat storing cost optimization method and device based on electric power big data | |
| CN112040002B (en) | Data fusion method, device and equipment based on power distribution cloud platform | |
| WO2024040845A1 (en) | Combine-harvester big data reconstruction method, and apparatus | |
| CN113824575B (en) | Method and device for identifying fault node, computing equipment and computer storage medium | |
| CN116109988B (en) | Anomaly monitoring method and system based on artificial intelligence and unmanned aerial vehicle | |
| CN114969163B (en) | Equipment operation and maintenance method and system based on big data | |
| CN107844868B (en) | Load prediction system fusing PLC technology and principal component analysis-BP neural network | |
| CN114595788B (en) | Transformer fault diagnosis method, device and equipment | |
| CN108830422B (en) | Intelligent driving optimization method, device and system | |
| US20240044988A1 (en) | Abnormality detection device, abnormality detection method, and computer program | |
| CN111859802B (en) | Method, system and equipment for calculating steady-state temperature of overhead conductor | |
| CN114912353A (en) | Wind direction correlation-based wind power plant wind direction numerical value calculation method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information |
Address after: Floor 31, Asia Pacific business building, No. 2 Hanzhong Road, Gulou District, Nanjing, Jiangsu 210005 Applicant after: LUCULENT SMART TECHNOLOGIES CO.,LTD. Address before: Room a808, World Trade Center building, 67 Shanxi Road, Gulou District, Nanjing, Jiangsu 210000 Applicant before: LUCULENT SMART TECHNOLOGIES CO.,LTD. |
|
| CB02 | Change of applicant information | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210423 |
|
| RJ01 | Rejection of invention patent application after publication |