A Quantization Method for Time-Series Data Compression in Non-Volatile Memory
A time-series data, non-volatile technology, applied in the direction of code conversion, electrical components, input/output to record carrier, etc., can solve the problem of voice recognition accuracy reduction, unbearable memory space consumption of dynamic time warping algorithm, memory space There are problems such as limitations to achieve the effect of reducing the amount of written data
Active Publication Date: 2020-05-26
CHONGQING UNIV
View PDF4 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Although it is possible to reduce the sampling rate of the time series (i.e. downsampling) to reduce the size of the reference database, this will reduce the accuracy of speech recognition
Since the non-volatile memory memory space is usually limited, the memory space consumption of the dynamic time warping algorithm (DTW) cannot be afforded
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment
[0052] The data used in this embodiment is an electrocardiogram time series data dataset, see Yanping Chen, EamonnKeogh, Bing Hu, Nurjahan Begum, Anthony Bagnall, Abdullah Mueen and GustavoBatista (2015). The UCR Time Series Classification Archive. URL (UCR time series data Category Archives) www.cs.ucr.edu / ~eamonn / time_series_data / ;
[0053] This time series data set includes the electrical activity of the human heartbeat over a period of time, with a total of 300 time series data, and each time series data represents a heartbeat cycle. In this embodiment, the first heartbeat period of the data set is taken as an example for illustration, and this embodiment needs to be repeated 300 times for the entire data set.
[0054] In step S01, select the smoothing window length ω=1, from the heartbeat cycle data S=(d 1 , d 2 … d n ) starts to execute step S02; wherein, n=140, that is, there are 140 heartbeat sampling data in S;
[0055] In step S02, by formula (1), the smoothing v...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a method for quantifying compression of time series data in nonvolatile memories. The method comprises the following steps of: 1, smoothing to-be-compressed time series data Sto obtain smoothed time series data (as shown in the specification), extracting a feature point in the smoothed time series data, and recording a value of a corresponding position in the original timeseries data S; 2, carrying out time series reconstruction by utilizing a linear interpolation method so as to obtain a reconstructed time series T; and 3, comparing the series T with series S by utilizing a dynamic time distortion algorithm, solving a distance between the two series, setting a distance threshold value, if the distance between the two time series is smaller than the threshold value, considering that the two time series are similar, and otherwise, considering that the two time series are not similar and re-adjusting the width of a smooth window. The method has the technical effects of realizing the rapid reconstruction of time series data when the time series data is stored in nonvolatile memories, and decreasing written data sizes of the nonvolatile memories while ensuringthat differences between the compressed data and the original data are located in an acceptable range.
Description
technical field [0001] The invention belongs to the technical field of data storage, and in particular relates to time series data compression of a nonvolatile memory. Background technique [0002] With the development of semiconductor technology, non-volatile memory (NVM), including phase-change memory (PCM) and memristor (memristor), is receiving more and more attention. The RAM produced by these new NVMs includes PCRAM, STT-RAM and RRAM, etc. The data stored in NVM has the characteristic of not being lost when power is turned off, that is, non-volatile. Compared with DRAM, NVM has a limited erasure cycle. In order to prolong the service life of NVM, existing technologies can be divided into two categories: reducing writes and wear leveling; reducing writes to NVM can prolong service life. [0003] Currently, time series data are generated in many application areas, such as wireless sensor networks, monitoring systems, and IoT scenarios. At the same time, in many applic...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06F3/06H03M7/30
Inventor 刘铎顾艺黃柏鈞李星妮
Owner CHONGQING UNIV