CN110210054A - A kind of sampled data preprocess method - Google Patents
A kind of sampled data preprocess method Download PDFInfo
- Publication number
- CN110210054A CN110210054A CN201910337138.3A CN201910337138A CN110210054A CN 110210054 A CN110210054 A CN 110210054A CN 201910337138 A CN201910337138 A CN 201910337138A CN 110210054 A CN110210054 A CN 110210054A
- Authority
- CN
- China
- Prior art keywords
- data
- sampled data
- sampled
- preprocess method
- 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
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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to a kind of sampled data preprocess methods, which comprises step S11: for the pretreatment of the missing values of sampled data；Step S12: it is pre-processed for high dimensional data.The present invention is the memory requirement for meeting blower unbalanced fault recovery process response simulation curve result data, the present invention provides combine History file data library and real-time internal memory database technology, quick-searching simulation curve library implementation method towards mass data, the characteristics of due to its mass data storage, high speed retrieval data, the mass memory of optimal Simulation result data may be implemented.
Description
[technical field]
The invention belongs to data simulation field more particularly to a kind of sampled data preprocess methods.
[background technique]
With the appearance of interconnection of large-scale power grids, low-frequency oscillation increases China's electric system year by year, and seriousness is even more than
Transient stability, becomes the major obstacle of system safe and stable operation.World power industrial development experience have shown that: with
Electric system scale constantly expand, the networking of big section and high-gain field regulator are widely used, so that jeopardizing the whole network
Multiple trend will be presented in the low-frequency oscillation of safety.The small signal stability of low-frequency oscillation and system is closely related, shows as
Opposite between generator amature to wave, power persistent oscillation on power transmission line, easily initiation large area blackout greatly threaten
The safe operation of system.It emulates particularly important for the protection in advance of system safety.Processing to emulation data is therein heavy
Want a ring.Therefore, a kind of new sampled data preprocess method is needed, the present invention is to meet blower unbalanced fault recovery process
The memory requirement for responding simulation curve result data, provides and combines History file data library and real-time internal memory database technology
, the quick-searching simulation curve library implementation method towards mass data, since its mass data storage, high speed retrieve data
The mass memory of optimal Simulation result data may be implemented in feature.
[summary of the invention]
In order to solve the above problem in the prior art, the invention proposes a kind of sampled data preprocess method, the party
Method includes:
Step S11: for the pretreatment of the missing values of sampled data；
Step S12: it is pre-processed for high dimensional data；It is specific: data high dimensional data is handled based on principal component analysis.
Further, the step S11 specifically: change curve is constructed based on sampled data, then utilizes the curve
Calculate missing values；Specifically: x at three sampled points of selection0, x1, x2Sampled value y0, y1, y2, three interpolation base letters are set
Number:Based on change curve
l2(x)=y0l0(x)+y1l1(x)+y2l2(x) Filling power is calculated；Wherein x is the sampling time；The x and x0, x1, x2Difference take phase
The difference in the sampling time of corresponding sampled point；Missing values are filled using the Filling power.
Further, the identical mode of sampling time interval selects sampled point.
Further, the selection of sampled point is carried out based on filling history.
Further, the various types of emulation data results generated in the blower unbalanced fault recovery process include
Corresponding relationship between sampled point and sampled data.
Further, step when inquiring v > v ', comprising:
Step A: using the intersection point of point inquiry seeking time sequence and straight line v=v ', acquiring time point t is t1′,t2′,…,
tk；
Step B: when k is even number.If t=t1' when time series be incremented by, then the period of v > v ' be [t1′,t2′],
[t3′,t4′],…,[tk-1′,tk′]；If t=t1' when time series successively decrease, period of v > v ' is [tmin′,t1′],[t2′,
t3′],[t4′,t5′],…,[tk-2′,tk-1′],[tk′,tmax'], wherein tmin' be time series initial state time, tmax′
For the time of final state.
Further, the Aggregation Query, specifically: inquiry sampled data reaches the critical values such as maximum value or minimum value
When corresponding sampling time or sampling timeframe.
It is further, described that data high dimensional data is handled based on principal component analysis, specifically: if being directed to sampled data X,
If sampled data X is the data matrix of n*m, in which: n is number of samples, and m is variable number；Then sampled data X is decomposed into
M=2 is arranged, then in the sum of the apposition of m vectorWherein: ti∈RnFor score
Vector, referred to as pivot, pi∈RmFor load vector；Pivot contribution rate is setIf the accumulation of preceding k pivot is contributed
Rate be more than before principal component threshold value is considered as k pivot reflect main information, described in setting
Further, the score vector is preset value；Principal component threshold value is equal to 85%.
Further, there is no unit, only multiple proportion after Interpolation-Radix-Function is reduced.
The beneficial effect comprise that responding simulation curve result data to meet blower unbalanced fault recovery process
Memory requirement, provide and combine History file data library and real-time internal memory database technology, towards the fast of mass data
The characteristics of speed retrieves simulation curve library implementation method, retrieves data due to its mass data storage, high speed, may be implemented optimal
The mass memory of Simulation result data.
[Detailed description of the invention]
Described herein the drawings are intended to provide a further understanding of the invention, constitutes part of this application, but
It does not constitute improper limitations of the present invention, in the accompanying drawings:
Fig. 1 is the schematic diagram of the compression method of time tag of the invention.
Fig. 2 is the step schematic diagram of magnitude compression method of the invention.
Fig. 3 is the magnitude compression schematic diagram of floating point values and cumulant in analog quantity of the invention.
Fig. 4 is blower unbalanced fault recovery process response curve library implementation method schematic diagram of the invention.
Fig. 5 is sampled data preprocess method schematic diagram of the invention.
[specific embodiment]
Come that the present invention will be described in detail below in conjunction with attached drawing and specific embodiment, illustrative examples therein and says
It is bright to be only used to explain the present invention but not as a limitation of the invention.
As shown in Figure 1, to a kind of blower unbalanced fault recovery process response curve library realization side applied by the present invention
Method is described in detail, by be directed to curve library data the characteristics of carry out data prediction, data compression, data query so that
The various types of emulation data results generated in blower unbalanced fault recovery process obtain efficiently store and quickly
Inquiry；
Described method includes following steps:
Step S1: data prediction；It is specific: to obtain sampled point, and the sampled data obtained to sampled point is located in advance
Reason；Shown in attached drawing 5, the data preprocessing method includes:
Step S11: for the pretreatment of the missing values of sampled data；Specifically: it is bent that variation is constructed based on sampled data
Then line calculates missing values using the curve；Specifically: x at three sampled points of selection0, x1, x2Sampled value y0, y1, y2,
Three Interpolation-Radix-Functions are set:
Based on change curve l2(x)=y0l0(x)+y1l1(x)+y2l2(x) Filling power is calculated；Wherein x is the sampling time；The x and x0,
x1, x2Difference take corresponding sampled point sampling time difference；Missing values are filled using the Filling power.
Preferred: the identical mode of sampling time interval selects sampled point；
It is preferred: the selection of sampled point is carried out based on filling history；
The various types of emulation data results generated in the blower unbalanced fault recovery process include that sampled point (is adopted
The sample time) and sampled data (sampled data value) between corresponding relationship；
Preferred: Interpolation-Radix-Function does not have unit, only multiple proportion after being reduced.
It is preferred: setting saturation limitation, it is described to be saturated the hits for being limited to the previous sampling time point of missing time points
According to the limitation of the difference of the sampled data with latter sampling time point；Wherein: the sampled data saves in the database；If lacked
The absolute value of the difference of the sampled data of mistake value and previous sampling time point is limited greater than the saturation, then will use the previous sampling time
Sampled data add or subtract the numerical value that saturation limit obtains to fill the missing values；
Step S12: it is pre-processed for high dimensional data；It is specific: data high dimensional data is handled based on principal component analysis；
It is described that data high dimensional data is handled based on principal component analysis, specifically: if being directed to sampled data X, if sampling
Data X is the data matrix of n*m, in which: n is number of samples, and m is variable number；Sampled data X is then decomposed into m vector
The sum of apposition, m=2 is set, thenWherein: ti∈RnFor score vector, claim
For pivot, pi∈RmFor load vector；Pivot contribution rate is setIf the accumulation contribution rate of preceding k pivot is more than master
K pivot reflects main information before ingredient threshold value is considered as, described in setting
Preferred: the score vector is preset value；Principal component threshold value is equal to 85%；
The data compression of step 2 progress Various types of data；It is specific: to carry out the compression of time tag, numerical value；
Step S21: carrying out the compression of time tag, specific: a reference value of storage time label, when by adjacent two
Between label value subtract each other；Judge whether former and later two differences are equal, it is such as equal, then equal times are increased by one, continues to read in next
A time tag, otherwise, the compression of the time tag data terminate, and export the difference equal times in this section of time tag data
Difference corresponding with the equal times；In view of in the emulation data result of blower unbalanced fault recovery process, the time
Label is that temporally the period, the difference redundancy of time tag was high in arithmetic progression；Therefore, the time can be directed to
The characteristics of label, is compressed；
Step S22: carrying out the compression of numerical value, specific: the numerical value include the numerical value of switching value, analog quantity numerical value and
The numerical value of cumulant；
Preferred: the numerical value is blower unbalanced fault recovery process response curve numerical value；
Step S221: the compression of switching value；Specifically: the value of switching value only there are two types of, or be 0 or be 1, switching value used
1 byte is come, if switching value is 0, is written one 0 to switching value memory block；If switching value is 1, be written to compressed file
One 1；
Preferred: the switching value is stored in compressed file；
Step S222: the compression of percentage amounts in analog quantity；Specifically: when the absolute value of percentage amounts difference is less than 256, adopt
It is compressed with LZW compression algorithm, when the absolute value of percentage amounts difference is greater than 256 difference, is added into dictionary, and defeated
Identifier out is identified the data and is compressed using LZ78 algorithm；
It is preferred: in order to decompression when can restoring data, need to open up a room be used to recording difference just
It is negative, herein, which is located to the most beginning of array, if difference is positive, is written 10, if difference is negative, is written 1
1；
As shown in Fig. 2, the compression method of the percentage amounts includes that steps are as follows:
Step 2221: initialization compression system dictionary reads in character；Wherein: fixed length needed for dictionary refers to lzw algorithm
String table, purposes are 8 codings that the character of input is mapped to fixed length, the i.e. index of dictionary；Character refers to that this processing is defeated
The character entered；
Step 2222: judge whether character length is greater than 256, is transferred to step 2223 if character length is no more than 256, it is no
Then it is transferred to step 2224；
Step 2223: character being connect after current string, current string, which refers to, to be made of processed character
Character string；
Step 2224: judging that character whether in dictionary, step 2225 is transferred to if not in dictionary, is otherwise transferred to step
Rapid 2226；
Step 2225: compressed file is written into character and its identifier, dictionary then is added in character and by the rope of dictionary
Draw as current string.The identifier of character is generated by LZ78 algorithm；
Step 2226: taking out index of the character in dictionary, and connect after current string；
Step 2227: search current string, if not being transferred to step 2228 if, is otherwise transferred to step whether in dictionary
2229；
Step 2228: the prefix character string of current string being written into file, and by the last character of current string
Symbol is used as current string.Prefix character string refers to the current string before character is added；
Step 2229: using the dictionary index of current string as current string；
Step 22210: reading in new character, and return step 2222；If not new character, terminates；
Preferred: the step 222 further includes step 223: carrying out the magnitude compression of floating point values and cumulant in analog quantity；
As shown in Fig. 3, the magnitude compression for carrying out floating point values and cumulant in analog quantity, includes the following steps:
Step 2231: initialization dictionary；Wherein, fixed-length string table needed for dictionary refers to lzw algorithm, purposes is will be defeated
The character entered is mapped to 8 codings of fixed length, the i.e. index of dictionary；
Step 2232: reading in numerical value, and it is pre-processed accordingly；In data prediction link, due to floating point values
Different from the data characteristics of the numerical value of cumulant, the preprocess method of use is also different；The numerical value of cumulant is by analog quantity
Numerical value is obtained according to regular hour periodic accumulation, therefore is pre-processed using differential technique to the numerical value of cumulant, that is, is used
The numerical value of current cumulant subtracts the numerical value of previous cumulant, compresses to obtained difference.The data of floating point values have
Preferable continuity, the fuctuation within a narrow range between adjacent two data is relatively common, and fluctuation is less, takes difference to its adjacent data,
This difference is usually similar, therefore using converting its percentage in range ability for floating point values, then to conversion after
Floating point values take the method for difference to be pre-processed；
Step 2233: reading in pretreated numerical value by byte and be converted into character string；
Step 2234: the character string of reading is added in current string；Current string refers to by processed word
Accord with the character string of composition；
Step 2235: whether search current string be not transferred to step 2236 then, is otherwise being transferred to step in dictionary
2237；
Step 2236: the dictionary index of the current maximum matching string of output.Maximum matching string refers to and longest in current string
The character string in dictionary that continuation character sequence matches；
Step 2237: dictionary is added in current string, using character string as current string；
Step 2238: judging whether numerical value compresses and finish, do not compress and finish then return step 2233, otherwise return step
2232；
Step 3 carries out data query；Specifically: support point inquiry, range query, Aggregation Query and/or similar inquiry；Institute
Inquiry is stated as the inquiry based on sampling time index and sampled data index；
The point inquiry, specifically: the value for inquiring sampled data corresponding sampling time when being particular value；Such as: inquiry
When a certain voltage value reaches 242kV.The point inquiry comprises the following steps:
Step A. finds the sampling time section comprising sampled data v ' using binary search in the dictionary of subsequence section
Ri；
Step B. seeks sampling time section RiDull subsequence array anch (i) in dull subsequence and straight line v=v '
Intersection point；For a dull subsequence C in anch (i)j, found out using binary search and CjThe intersection point of straight line v=v ', obtains
To sampling time t；
The range query, specifically: compressed file is inquired to obtain its corresponding sampled data model according to the sampling time
It encloses；
It is described according to the sampling time inquire compressed file to obtain its corresponding sampled data range, specifically: inquiry adopt
When sample time t within which period, the value v of sampled data is greater than v ', or is less than v ', or between v ' and v "；Wherein: v ' with
V " is inquiry boundary；For example, its value of which of a certain voltage is more than 242kV time, or between 220kV to 242kV；Due to entire
Time series is continuous subsequence, as long as finding out the intersection point of time series Yu straight line v=v ', so that it may which output meets condition
Period；Step when inquiry v > v ' is given below, other inquiries are similar；
Step A: using the intersection point of point inquiry seeking time sequence and straight line v=v ', acquiring time point t is t1′,t2′,…,
tk；
Step B: when k is even number.If t=t1' when time series be incremented by, then the period of v > v ' be [t1′,t2′],
[t3′,t4′],…,[tk-1′,tk′]；If t=t1' when time series successively decrease, period of v > v ' is [tmin′,t1′],[t2′,
t3′],[t4′,t5′],…,[tk-2′,tk-1′],[tk′,tmax'], wherein tmin' be time series initial state time, tmax′
For the time of final state.When k is odd number, method is similar；
The Aggregation Query, specifically: inquiry sampled data reaches corresponding when the critical values such as maximum value or minimum value
Sampling time or sampling timeframe；Aggregation Query mainly includes t when time series reaches maximum value or minimum value etc.,
Such as when certain voltage value is worth and to reach highest on July 16th, 2005；
For the time t for reaching maximum value in the hope of time series below, illustrate to realize that step, other query steps are similar；
Step A: the maximum section R of v value is found in the dictionary of subsequence sectionm, RmUpper bound vmaxExactly time series
Maximum value；
Step B: using the method for point inquiry, R is soughtmInterior dull subsequence and straight line v=vmaxIntersection point, these intersection points
T value be required；
The similar inquiry, specifically: inquiry sampled data identical sampling time or time range sequence；The phase
Like inquiring primarily directed to after certain curve to be checked, or a few constraint conditions of proposition, searched in simulation curve library all
Meet the curve of condition, and exports result；Include the following steps:
Step A: being directed to curve to be checked, or the constraint condition proposed, seeks the higher crucial subsequence section of its weight
(the sum of its weight is more than 85% i.e. it is believed that these crucial subsequence sections reflect the main information of required curve), and
According to weight sequencing；
Step B: it chooses the highest crucial subsequence section of weight and is searched in the dictionary of subsequence section into binary, find this
All qualified curves are found by being directed toward the pointer of frequency behind subsequence section, the key area for then selecting weight time high
Between continue to search in the result；
Step C: circulation carries out step B, finishes until by all crucial subsequence range lookups selected, then remaining song
Line is required；
In view of blower unbalanced fault recovery process responds in Simulation result data analysis, for the master of data retrieval
It wants demand to be that quickly and finds several monitor curves for meeting search condition, and these curve rapidly extractings are come out.
Therefore the introducing of the mode of inverted index value inquiry can be used to for search condition to be decomposed into inquiry, range query and an aggregation
The combination of inquiry to quickly extract the curve data of caching from real-time internal memory database, and then arrives History file data
The more information of the curve is found in library so that program is further analyzed；
In several embodiments provided by the present invention, it should be understood that disclosed method and terminal can pass through it
Its mode is realized.For example, the apparatus embodiments described above are merely exemplary, for example, the division of the module, only
Only a kind of logical function partition, there may be another division manner in actual implementation.
In addition, the technical solution in above-mentioned several embodiments can be combined with each other and replace in the case where not conflicting
It changes.
The module as illustrated by the separation member may or may not be physically separated, aobvious as module
The component shown may or may not be physical unit, it can and it is in one place, or may be distributed over multiple
In network unit.Some or all of the modules therein can be selected to realize the mesh of this embodiment scheme according to the actual needs
's.
It, can also be in addition, each functional module in each embodiment of the present invention can integrate in one processing unit
It is that each unit physically exists alone, can also be integrated in one unit with two or more units.Above-mentioned integrated list
Member both can take the form of hardware realization, can also realize in the form of hardware adds software function module.
It is obvious to a person skilled in the art that invention is not limited to the details of the above exemplary embodiments, Er Qie
In the case where without departing substantially from spirit or essential attributes of the invention, the present invention can be realized in other specific forms.Therefore, no matter
From the point of view of which point, the present embodiments are to be considered as illustrative and not restrictive, and the scope of the present invention is by appended power
Benefit requires rather than above description limits, it is intended that all by what is fallen within the meaning and scope of the equivalent elements of the claims
Variation is included in the present invention.Any attached associated diagram label in claim should not be considered as right involved in limitation to want
It asks.Furthermore, it is to be understood that one word of " comprising " does not exclude other units or steps, odd number is not excluded for plural number.It is stated in system claims
Multiple modules or device can also be implemented through software or hardware by a module or device.The first, the second equal words
It is used to indicate names, and does not indicate any particular order.
Finally it should be noted that the above examples are only used to illustrate the technical scheme of the present invention and are not limiting, although reference
Preferred embodiment describes the invention in detail, those skilled in the art should understand that, it can be to of the invention
Technical solution is modified or equivalent replacement, without departing from the spirit and scope of the technical solution of the present invention.
Claims (10)
1. a kind of sampled data preprocess method characterized by comprising
Step S11: for the pretreatment of the missing values of sampled data；
Step S12: it is pre-processed for high dimensional data；It is specific: data high dimensional data is handled based on principal component analysis.
2. sampled data preprocess method according to claim 1, which is characterized in that the step S11 specifically: be based on
Sampled data constructs change curve, then calculates missing values using the curve；Specifically: x at three sampled points of selection0,
x1, x2Sampled value y0, y1, y2, three Interpolation-Radix-Functions are set:Based on change curve l2
(x)=y0l0(x)+y1l1(x)+y2l2(x) Filling power is calculated；Wherein x is the sampling time；The x and x0, x1, x2Difference take relatively
The difference in the sampling time for the sampled point answered；Missing values are filled using the Filling power.
3. sampled data preprocess method according to claim 2, which is characterized in that the identical mode of sampling time interval
Select sampled point.
4. sampled data preprocess method according to claim 3, which is characterized in that carry out sampled point based on filling history
Selection.
5. sampled data preprocess method according to claim 4, which is characterized in that the blower unbalanced fault restores
The various types of emulation data results generated in the process include the corresponding relationship between sampled point and sampled data.
6. sampled data preprocess method according to claim 5, which is characterized in that step when inquiry v > v ', comprising:
Step A: using the intersection point of point inquiry seeking time sequence and straight line v=v ', acquiring time point t is t '1,t′2,…,t′k；
Step B: when k is even number.If t=t1' when time series be incremented by, then the period of v > v ' be [t1′,t2′],[t3′,
t4′],…,[tk-1′,tk′]；If t=t1' when time series successively decrease, period of v > v ' is [tmin′,t1′],[t2′,t3′],
[t4′,t5′],…,[tk-2′,tk-1′],[tk′,tmax'], wherein tmin' be time series initial state time, tmax' for most
The time of state afterwards.
7. sampled data preprocess method according to claim 6, which is characterized in that the Aggregation Query, specifically: it looks into
It askes sampled data and reaches corresponding sampling time or sampling timeframe when the critical values such as maximum value or minimum value.
8. sampled data preprocess method according to claim 7, which is characterized in that described to be handled based on principal component analysis
Data high dimensional data, specifically: if sampled data X is directed to, if sampled data X is the data matrix of n*m, in which: n is sample
This number, m are variable number；Sampled data X is then decomposed into the sum of the apposition of m vector, m=2 is set, thenWherein: ti∈RnFor score vector, referred to as pivot, pi∈RmFor load to
Amount；Pivot contribution rate is setIf the accumulation contribution rate of preceding k pivot is more than k before principal component threshold value is considered as
Pivot reflects main information, described in setting
9. sampled data preprocess method according to claim 8, which is characterized in that the score vector is preset value；
Principal component threshold value is equal to 85%.
10. sampled data preprocess method according to claim 9, which is characterized in that after Interpolation-Radix-Function is reduced
There is no unit, only multiple proportion.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910337138.3A CN110210054A (en) | 2019-04-25 | 2019-04-25 | A kind of sampled data preprocess method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910337138.3A CN110210054A (en) | 2019-04-25 | 2019-04-25 | A kind of sampled data preprocess method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110210054A true CN110210054A (en) | 2019-09-06 |
Family
ID=67786353
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910337138.3A Pending CN110210054A (en) | 2019-04-25 | 2019-04-25 | A kind of sampled data preprocess method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110210054A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111046027A (en) * | 2019-11-25 | 2020-04-21 | 北京百度网讯科技有限公司 | Missing value filling method and device for time series data |
| CN114726380A (en) * | 2022-06-07 | 2022-07-08 | 西南交通大学 | Monitoring data lossless compression method, device, equipment and readable storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106529079A (en) * | 2016-11-29 | 2017-03-22 | 上海电机学院 | Chemical process failure detection method based on failure-dependent principal component space |
| CN109362083A (en) * | 2018-10-31 | 2019-02-19 | 北京交通大学 | The building of high-speed railway wireless channel database and data preprocessing method |
| CN109376953A (en) * | 2018-11-21 | 2019-02-22 | 国网冀北电力有限公司 | A kind of Middle and long term electricity consumption forecasting method and system |
| CN109508811A (en) * | 2018-09-30 | 2019-03-22 | 中冶华天工程技术有限公司 | Parameter prediction method is discharged based on principal component analysis and the sewage treatment of shot and long term memory network |
-
2019
- 2019-04-25 CN CN201910337138.3A patent/CN110210054A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106529079A (en) * | 2016-11-29 | 2017-03-22 | 上海电机学院 | Chemical process failure detection method based on failure-dependent principal component space |
| CN109508811A (en) * | 2018-09-30 | 2019-03-22 | 中冶华天工程技术有限公司 | Parameter prediction method is discharged based on principal component analysis and the sewage treatment of shot and long term memory network |
| CN109362083A (en) * | 2018-10-31 | 2019-02-19 | 北京交通大学 | The building of high-speed railway wireless channel database and data preprocessing method |
| CN109376953A (en) * | 2018-11-21 | 2019-02-22 | 国网冀北电力有限公司 | A kind of Middle and long term electricity consumption forecasting method and system |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111046027A (en) * | 2019-11-25 | 2020-04-21 | 北京百度网讯科技有限公司 | Missing value filling method and device for time series data |
| CN114726380A (en) * | 2022-06-07 | 2022-07-08 | 西南交通大学 | Monitoring data lossless compression method, device, equipment and readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Gueniche et al. | Compact prediction tree: A lossless model for accurate sequence prediction | |
| CN110210054A (en) | A kind of sampled data preprocess method | |
| CN112699605B (en) | Charging pile fault element prediction method and system | |
| Tan et al. | System identification and linearisation using genetic algorithms with simulated annealing | |
| US11620260B2 (en) | Record property synchronization in a network computing system | |
| CN103036697B (en) | Multi-dimensional data duplicate removal method and system | |
| CN112953550A (en) | Data compression method, electronic device and storage medium | |
| EP0127815B1 (en) | Data compression method | |
| CN111241244A (en) | Big data-based answer position acquisition method, device, equipment and medium | |
| CN109597574B (en) | Distributed data storage method, server and readable storage medium | |
| CN109325089A (en) | A kind of non-pointing object querying method, device, terminal device and storage medium | |
| CN111654392A (en) | Low-voltage distribution network topology identification method and system based on mutual information | |
| CN110209696A (en) | A kind of blower unbalanced fault recovery process response curve library implementation method | |
| CN104077272B (en) | A kind of method and apparatus of dictionary compression | |
| CN112528650B (en) | Bert model pre-training method, system and computer equipment | |
| CN111078652A (en) | Filing and compressing method and device for logistics box codes | |
| Chen et al. | Using difficulty of prediction to decrease computation: Fast sort, priority queue and convex hull on entropy bounded inputs | |
| CN112181302A (en) | Data multilevel storage and access method and system | |
| CN111143695A (en) | Searching method, searching device, server and storage medium | |
| CN113536508B (en) | Manufacturing network node classification method and system | |
| Vishwakarma et al. | Order reduction using the advantages of differentiation method and factor division algorithm | |
| CN114780536A (en) | SQL Server database index creation method and device, electronic equipment and storage medium | |
| JP3346626B2 (en) | Data compression device | |
| CN112801817A (en) | Electric energy quality data center construction method and system | |
| KR101793005B1 (en) | Incremental high utility pattern mining method with static and dynamic databases |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190906 |