CN105404772A - Segmented system memory effect based adaptive historical data analysis method - Google Patents
Segmented system memory effect based adaptive historical data analysis method Download PDFInfo
- Publication number
- CN105404772A CN105404772A CN201510746431.7A CN201510746431A CN105404772A CN 105404772 A CN105404772 A CN 105404772A CN 201510746431 A CN201510746431 A CN 201510746431A CN 105404772 A CN105404772 A CN 105404772A
- Authority
- CN
- China
- Prior art keywords
- response
- weighting coefficient
- time slice
- memory effect
- system memory
- 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
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16Z—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS, NOT OTHERWISE PROVIDED FOR
- G16Z99/00—Subject matter not provided for in other main groups of this subclass
Abstract
The invention provides a segmented system memory effect based adaptive historical data analysis method. The method comprises the following steps of: S1, dividing the time of system data into a plurality of time segments and setting corresponding weighting coefficients; S2, according to a time node and a weighting coefficient of each time segment, calculating a prediction response; S3, obtaining a practical response of a system and calculating a response error according to the prediction response and the practical response; S4, determining whether the response error is greater than a preset threshold; S5, if the response error is greater than the preset threshold, determining whether the repeating frequency of the step S2 is higher than a preset frequency; S6, if the repeating frequency of the step S2 is lower than the preset frequency, resetting a corresponding weighting coefficient and repeatedly performing the step S2; and S7, if the repeating frequency of the step S2 is higher than the preset frequency, resetting a preset segmentation rule and a corresponding weighting coefficient, and repeatedly performing the step S2. According to the method, the overhead of computing resources and computing time is reduced while a computing result is ensured to be accurate.
Description
Technical field
The invention belongs to data analysis field, be specifically related to a kind of adaptive history data analysing method based on segmented system memory effect.
Background technology
In the process of industrial data analysis and excavation, need to choose data in the past and substitute into and calculate, draw the data result that user needs.Certainly, the data chosen are more, and the result of calculating can more close to actual value; If but simply all data in past were all substituted into calculating, the expense of computational resource and computing time could be very huge, and the contradiction between this result of calculation accuracy and computational resource expense constrains the development of industrial data analysis and excavation.
Summary of the invention
The object of the invention is to for the deficiencies in the prior art, provide a kind of adaptive history data analysing method based on segmented system memory effect, the method can solve the problem that result of calculation accuracy and computational resource expense exist contradiction well.
For reaching above-mentioned requirements, the technical scheme that the present invention takes is: provide a kind of adaptive history data analysing method based on segmented system memory effect, comprise the following steps:
S1, according to preset chopping rule the time of system data is divided into multiple time slice, and for each time slice setting correspondence weighting coefficient;
S2, according to the timing node of each time slice and weighting coefficient computational prediction response corresponding to each time slice;
S3, obtain phylogenetic real response, according to this predicated response and this real response calculated response error;
S4, judge whether response error is greater than predetermined threshold value;
If S5 is greater than predetermined threshold value, whether the multiplicity of determining step S2 is greater than preset times;
If S6 is less than preset times, calculates and simulate weighting coefficient correcting value, is that each time slice resets corresponding weighting coefficient according to weighting coefficient correcting value, and repeated execution of steps S2;
If S7 is greater than preset times, reset default chopping rule, according to default chopping rule, the time period of system data is reclassified as multiple time slice, and reset corresponding weighting coefficient for each time slice, and repeated execution of steps S2.
Compared with prior art, the present invention has the following advantages: data are temporally carried out segmentation, and be that each time slice sets suitable weighting coefficient according to the correlativity of data, the data only choosing correlativity high when adding up calculate, make result of calculation more accurate, decrease the expense of computational resource and computing time simultaneously.
Accompanying drawing explanation
Accompanying drawing described herein is used to provide further understanding of the present application, form a application's part, use identical reference number to represent same or analogous part in the drawings, the schematic description and description of the application, for explaining the application, does not form the improper restriction to the application.In the accompanying drawings:
Fig. 1 shows process flow diagram of the present invention.
Embodiment
For making the object of the application, technical scheme and advantage clearly, below in conjunction with drawings and the specific embodiments, the application is described in further detail.For the sake of simplicity, eliminate in below describing and well known to a person skilled in the art some technical characteristic.
According to one embodiment of present invention, a kind of adaptive history data analysing method based on segmented system memory effect is provided, as shown in Figure 1, comprises the following steps:
S1, according to default chopping rule, the time of system data is divided into multiple time slice, and is weighting coefficient corresponding to each time slice setting;
Such as to predict the supplies consumption situation in October, the system data in existing July, August, September, default chopping rule is that system data is carried out segmentation by every two weeks, is divided into 6 sections, and the weighting coefficient of setting is followed successively by 0.1,0.1,0.1,0.2,0.2,0.3.
S2, according to the timing node of each time slice and weighting coefficient computational prediction response corresponding to each time slice, the computing formula of predicated response is
wherein, n represents the quantity of time slice, t
nrepresent the timing node of the n-th time slice, w
nrepresent that the weighting coefficient that the n-th time slice is corresponding, f (t) are system instantaneous function, H (t) is system prediction response function;
By upper example, the value of n is 6, and timing node is respectively July 1, July 15, August 1, August 15, September 1, September 15, September 30, weighting coefficient w
nbe respectively 0.1,0.1,0.1,0.2,0.2,0.3, these data substituted in the computing formula of predicated response, calculate predicated response.
S3, obtain phylogenetic real response, according to predicated response and real response calculated response error, response error is the absolute value of the difference of predicated response and real response;
S4, judge whether response error is greater than predetermined threshold value, this predetermined threshold value is set by the user, and can adjust according to actual conditions;
If S5 is greater than predetermined threshold value, whether the multiplicity of determining step S2 is greater than preset times, and preset times is 30 times herein;
If S6 is less than preset times, calculate and simulate weighting coefficient correcting value, if response error and predetermined threshold value deviation are comparatively large, then weighting coefficient correcting value is also larger, be that each time slice resets corresponding weighting coefficient according to this weighting coefficient correcting value, and repeated execution of steps S2;
If S7 is greater than preset times, if namely the multiplicity of step S2 is more than 30 times, then show it may is that chopping rule is not too accurate, now need to reset default chopping rule, according to default chopping rule, the time period of system data is reclassified as multiple time slice, such as carry out time slice by each week, and reset corresponding weighting coefficient for each time slice, and repeated execution of steps S2.
According to one embodiment of present invention, the method also comprises:
If S8 is less than predetermined threshold value, illustrate that the deviation of predicated response and real response is in range of receiving, the default chopping rule then now drawn and weighting coefficient are default chopping rule best in the supplies consumption situation in prediction October and weighting coefficient, carry out subsequent treatment to this default chopping rule and weighting coefficient.
The present invention is in the systematic training stage, step S1 is by manually carrying out, namely initial default chopping rule and corresponding initial weighting coefficient is inputted according to the experience of user, then by computer executed step S2, S3, S4, S5, weighting coefficient correcting value in step S6 is calculated and matching by artificial, and reset weighting coefficient according to this weighting coefficient correcting value, repeated execution of steps 2 constantly adjusts default chopping rule and weighting coefficient, until response error is less than predetermined threshold value, the default chopping rule now drawn and weighting coefficient are for best, statistical study is carried out to the system data selected according to this default chopping rule and weighting coefficient and the data precision drawn is higher, and because the data only choosing correlativity high calculate, greatly reduce the expense of computational resource and computing time.After training stage of the present invention completes, user only needs to input the statistics and analysis that predetermined threshold value gets final product automatic completion system data in use.
The above embodiment only represents several embodiment of the present invention, and it describes comparatively concrete and detailed, but can not be interpreted as limitation of the scope of the invention.It should be pointed out that for the person of ordinary skill of the art, without departing from the inventive concept of the premise, can also make some distortion and improvement, these all belong to scope.Therefore protection scope of the present invention should be as the criterion with described claim.
Claims (5)
1., based on the adaptive history data analysing method of segmented system memory effect, it is characterized in that, comprise the following steps:
S1, according to preset chopping rule the time of system data is divided into multiple time slice, and for each time slice setting correspondence weighting coefficient;
S2, according to the timing node of each time slice and weighting coefficient computational prediction response corresponding to each time slice;
S3, obtain phylogenetic real response, according to described predicated response and described real response calculated response error;
S4, judge whether described response error is greater than predetermined threshold value;
If S5 is greater than predetermined threshold value, whether the multiplicity of determining step S2 is greater than preset times;
If S6 is less than preset times, calculates and simulate weighting coefficient correcting value, is that each time slice resets corresponding weighting coefficient according to described weighting coefficient correcting value, and repeated execution of steps S2;
If S7 is greater than preset times, reset default chopping rule, according to default chopping rule, the time period of system data is reclassified as multiple time slice, and reset corresponding weighting coefficient for each time slice, and repeated execution of steps S2.
2. the adaptive history data analysing method based on segmented system memory effect according to claim 1, is characterized in that: also comprise:
If S8 is less than predetermined threshold value, then carry out subsequent treatment.
3. the adaptive history data analysing method based on segmented system memory effect according to claim 1, is characterized in that: described preset times is 30 times.
4. the adaptive history data analysing method based on segmented system memory effect according to claim 1, is characterized in that: the computing formula of described predicated response is
wherein, n represents the quantity of time slice, t
nrepresent the timing node of the n-th time slice, w
nrepresent that the weighting coefficient that the n-th time slice is corresponding, f (t) are system instantaneous function, H (t) is system prediction response function.
5. the adaptive history data analysing method based on segmented system memory effect according to claim 1, is characterized in that: described response error is the absolute value of the difference of predicated response and real response.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510746431.7A CN105404772A (en) | 2015-11-04 | 2015-11-04 | Segmented system memory effect based adaptive historical data analysis method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510746431.7A CN105404772A (en) | 2015-11-04 | 2015-11-04 | Segmented system memory effect based adaptive historical data analysis method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105404772A true CN105404772A (en) | 2016-03-16 |
Family
ID=55470258
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510746431.7A Pending CN105404772A (en) | 2015-11-04 | 2015-11-04 | Segmented system memory effect based adaptive historical data analysis method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105404772A (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101438335A (en) * | 2006-03-03 | 2009-05-20 | 因瑞克斯有限公司 | Assessing road traffic conditions using data from mobile data sources |
| CN102568207A (en) * | 2012-02-02 | 2012-07-11 | 北京捷易联科技有限公司 | Traffic data processing method and device |
| CN102968670A (en) * | 2012-10-23 | 2013-03-13 | 北京京东世纪贸易有限公司 | Method and device for predicting data |
| US8731977B1 (en) * | 2013-03-15 | 2014-05-20 | Red Mountain Technologies, LLC | System and method for analyzing and using vehicle historical data |
-
2015
- 2015-11-04 CN CN201510746431.7A patent/CN105404772A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101438335A (en) * | 2006-03-03 | 2009-05-20 | 因瑞克斯有限公司 | Assessing road traffic conditions using data from mobile data sources |
| CN102568207A (en) * | 2012-02-02 | 2012-07-11 | 北京捷易联科技有限公司 | Traffic data processing method and device |
| CN102968670A (en) * | 2012-10-23 | 2013-03-13 | 北京京东世纪贸易有限公司 | Method and device for predicting data |
| US8731977B1 (en) * | 2013-03-15 | 2014-05-20 | Red Mountain Technologies, LLC | System and method for analyzing and using vehicle historical data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| TWI794157B (en) | Automatic multi-threshold feature filtering method and device | |
| Mei et al. | Signal processing on graphs: Estimating the structure of a graph | |
| CN111783953A (en) | 24-point power load value 7-day prediction method based on optimized LSTM network | |
| CN113469325B (en) | Hierarchical federation learning method for edge aggregation interval self-adaptive control, computer equipment and storage medium | |
| CN104766175A (en) | Power system abnormal data identifying and correcting method based on time series analysis | |
| US20230028574A1 (en) | Traffic prediction method, device, and storage medium | |
| CN106406830B (en) | Quasi-periodic signal processing method and device and computer readable storage medium | |
| CN107133190A (en) | The training method and training system of a kind of machine learning system | |
| CN112417752A (en) | Cloud layer track prediction method and system based on convolution LSTM neural network | |
| CN114819374A (en) | Regional new energy ultra-short term power prediction method and system | |
| CN109447336B (en) | Optimized control method for water level between upstream reservoir and reverse regulation reservoir dam thereof | |
| CN116523086A (en) | Single well production dynamic prediction method based on long-short-term memory depth neural network | |
| US10628541B2 (en) | Method and apparatus for selecting integrated circuit device neural network modeling sample | |
| CN111249106B (en) | Training control device and system of old people rehabilitation robot | |
| CN105404772A (en) | Segmented system memory effect based adaptive historical data analysis method | |
| CN109902266B (en) | Riverway flow calculation method based on Copula function | |
| CN107644233A (en) | FILTERSIM analogy methods based on Cluster Classification | |
| JP6186303B2 (en) | Traffic amount upper limit prediction apparatus, method and program | |
| CN109035178A (en) | A kind of multi-parameter value tuning method applied to image denoising | |
| CN103268614B (en) | A kind of for many prospects be divided into cut prospect spectrum drawing generating method | |
| CN112700050B (en) | Method and system for predicting ultra-short-term 1 st point power of photovoltaic power station | |
| CN108599976B (en) | Network quality evaluation model construction method and device and computer equipment | |
| CN110717277A (en) | Time sequence wind speed simulation method | |
| CN110929849A (en) | Neural network model compression method and device | |
| CN104010316A (en) | Method for telephone traffic prediction through cell-level wireless network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into 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: 20160316 |