久驰股票网-股票开户,股票基金入门知识 b转a:Matlab 2018a 比2017b有哪些改进?作者:知乎用户

b转a:Matlab 2018a 比2017b有哪些改进?作者:知乎用户

发布时间:2020-02-10 18:00   来源:网络 点击 :

写了不少R2018a的更新, 特别是信号处理方面的:

MATLAB



更新:

重点查看了一下自己感兴趣的工具箱:

信号处理工具箱:

• Signal Analyzer App: Preprocess signals by smoothing and filtering
• Signal Analyzer App: Detect transients and perform time-frequency analysis using
scalogram view
• One-Step Filtering: Filter signals using lowpass, highpass, bandpass, and bandstop
responses
• Spectral Measurements: Kurtogram, spectral kurtosis, and spectral entropy
• Rotating Machinery: Estimate and track rotational speed from vibration signals

1 Signal Analyzer App继续增加功能: 1 信号预处理, 包括平滑与滤波. 2 侦测瞬时变化和画尺度谱.(感觉是要添加小波分析?)

2 不太明白"One-Step Filtering"是什么意思, 是易用的意思?

3 Kurtogram, spectral kurtosis, and spectral entropy 这几个名词之前没听说过, 信号处理果然博大精深, 等新版本出来再学吧.


计量经济学工具箱:

• Econometric Modeler App: Perform time series analysis, specification testing,
modeling, and diagnostics

编程编多了后, 就越来越喜欢用App了, 先通过App, 来 get idea, 再编程, 可以节省不少时间.

小波工具箱:

Filter Banks for Continuous Wavelet Transform: Create, visualize and use filter
banks for Time-Frequency Analysis
• Filter Banks for Discrete Wavelet Transform: Explore time-frequency characteristics
of wavelets and scaling functions
• Constant-Q Transform: Perform adaptive time-frequency analysis using
nonstationary Gabor frames
• Scalogram View: Detect transients and perform time-frequency analysis with the
Signal Analyzer App
• Empirical Mode Decomposition and Hilbert-Huang Transform: Data-adaptive timefrequency
analysis for nonlinear and nonstationary processes
• Continuous Wavelet Transform: Improved control of time-frequency parameters and
timetable support

学习小波算是我个人学习计划吧. 有一个亮点: EMD分解希尔伯特-黄变换加入了MATLAB套餐中了!

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

大概看了一下"Release Notes", 最让我眼前一亮的是:

• Startup: Increased speed of MATLAB startup time
• Execution Engine: Execute tight loops with scalar math faster
• Execution Engine: Improved performance for common programming patterns

MATLAB最近几个版本不断在提速, 不更新版本的话, "MATLAB循环速度慢"这种观点很可能过时了.

我在使用的时候, 经常发现循环比矩阵运算快的例子, 预计以后这样的例子会越来越多.

举个例子(在R2017b上测试):

N = 1e2;
nloops = 10000;
x = randn(N, 1);
w = randn(N, 1);
tic;
for jj = 1:nloops
    s1 = w' * x;
end
t1 = toc;

tic;
for jj = 1:nloops
    s2 = 0;
    for ii = 1:length(x)
        s2 = s2 + w(ii) * x(ii);
    end
end
t2 = toc;

t1/t2
s1 - s2

结果:

ans =
   1.933815776861880
ans =
     3.552713678800501e-15

循环运算速度是矩阵运算速度的1.9倍.

当然, 将N改成更大, 比如1e5, 那就是矩阵运算快了.


"common programming patterns"不知道指的是什么, 等正式出来再看吧.

更多回复:

知乎网友崔星星:

根据 @芷菡 在Matlab R2017b中的测试,我分别Matlab R2017b和Matlab R2018a Prerelease测试结果如下:

2017b的:

ans =
   1.7083
ans =
  4.4409e-16

2018a的:

ans =
   4.1039
ans =
 -8.8818e-16

知乎网友张海峰:

Deep Learning
Plot and analyze your network using network analyzer, generate CUDA code that integrates with TensorRT, and deploy deep learning networks to Intel and ARM processors.

Data Analytics
Find outliers and smooth out-of-memory data using tall arrays, summarize timetables into time bins, and generate C code for ensemble and k-nearest neighbor classifiers using Kd-tree search.


Predictive Maintenance Toolbox
A new product for designing and testing condition monitoring and predictive maintenance algorithms.

Vehicle Dynamics Blockset
A new product for modeling and simulating vehicle dynamics in a virtual 3D environment.


知乎网友LOYA衍生品交易员:

挺好的,谢谢分享,前来学习。


Matlab 2018a 比2017b有哪些改进? - 知乎
zhihu.com/question/2643


知乎网友崔星星:

根据 @芷菡 在Matlab R2017b中的测试,我分别Matlab R2017b和Matlab R2018a Prerelease测试结果如下:

2017b的:

ans =
   1.7083
ans =
  4.4409e-16

2018a的:

ans =
   4.1039
ans =
 -8.8818e-16

猜你喜欢