一组数据,用vb求fft,数据需要归一化吗

归一化后求FFT和没有归一化的值幅值相差一个倍数，各次谐波的角度不变。这个倍数就是归一化的过程缩小的倍数。

FFT是快速傅立叶分解，FFT后出来的数据不是点，第一个结果是直流分量，其模需要除以采样点数，才能是幅值。如果是150点，出来的150个数据是0,1,2,.....75,74,73.....2,1。所以直流量没有对称数据。

数据标准化（归一化）处理是数据挖掘的一项棚散基础工作，不同评价指标往往具有不同的量纲和量纲单位，这样的情况会影响到数据分析的结链搜氏果，为了消除指标之间的量纲影响，需要进行数据标准化处理，以解决数据指标之间的可比性。原始数据经过数据标准化处理后，各指标处于同一漏桐数量级，适合进行综合对比评价。以下是常用的归一化方法：

min-max标准化（Min-Max Normalization）

也称为离差标准化，是对原始数据的线性变换，使结果值映射到[0 , 1]之间。转换函数如下：

其中max为样本数据的最大值，min为样本数据的最小值。这种方法有个缺陷就是当有新数据加入时，可能导致max和min的变化，需要重新定义。

这个哥们博客有很多相关的

这个好像可以进行FFT和IFFT

前几天刚考完数字信号处理 学的这个晕啊

呵呵

这让悉个是FFT的

*模块********************************************************

'FFT0 数组下标以0开始 FFT1 数组下标胡滑茄以1开始

'AR() 数据裤察实部 AI() 数据虚部

'N 数据点数，为2的整数次幂

'NI 变换方向 1为正变换，-1为反变换

'***************************************************************

Public Const Pi = 3.1415926

Public Function FFT0(AR() As Double, AI() As Double, N As Integer, ni As Integer)

Dim i As Integer, j As Integer, k As Integer, L As Integer, M As Integer

Dim IP As Integer, LE As Integer

Dim L1 As Integer, N1 As Integer, N2 As Integer

Dim SN As Double, TR As Double, TI As Double, WR As Double, WI As Double

Dim UR As Double, UI As Double, US As Double

M = NTOM(N)

N2 = N / 2

N1 = N - 1

SN = ni

j = 1

For i = 1 To N1

If i <j Then

TR = AR(j - 1)

AR(j - 1) = AR(i - 1)

AR(i - 1) = TR

TI = AI(j - 1)

AI(j - 1) = AI(i - 1)

AI(i - 1) = TI

End If

k = N2

While (k <j)

j = j - k

k = k / 2

Wend

j = j + k

Next i

For L = 1 To M

LE = 2 ^ L

L1 = LE / 2

UR = 1#

UI = 0#

WR = Cos(Pi / L1)

WI = SN * Sin(Pi / L1)

For j = 1 To L1

For i = j To N Step LE

IP = i + L1

TR = AR(IP - 1) * UR - AI(IP - 1) * UI

TI = AI(IP - 1) * UR + AR(IP - 1) * UI

AR(IP - 1) = AR(i - 1) - TR

AI(IP - 1) = AI(i - 1) - TI

AR(i - 1) = AR(i - 1) + TR

AI(i - 1) = AI(i - 1) + TI

Next i

US = UR

UR = US * WR - UI * WI

UI = UI * WR + US * WI

Next j

Next L

If SN <>-1 Then

For i = 1 To N

AR(i - 1) = AR(i - 1) / N

AI(i - 1) = AI(i - 1) / N

Next i

End If

End Function

Public Function FFT1(AR() As Double, AI() As Double, N As Integer, ni As Integer)

Dim i As Integer, j As Integer, k As Integer, L As Integer, M As Integer

Dim IP As Integer, LE As Integer

Dim L1 As Integer, N1 As Integer, N2 As Integer

Dim SN As Double, TR As Double, TI As Double, WR As Double, WI As Double

Dim UR As Double, UI As Double, US As Double

M = NTOM(N)

N2 = N / 2

N1 = N - 1

SN = ni

j = 1

For i = 1 To N1

If i <j Then

TR = AR(j)

AR(j) = AR(i)

AR(i) = TR

TI = AI(j)

AI(j) = AI(i)

AI(i) = TI

End If

k = N2

While (k <j)

j = j - k

k = k / 2

Wend

j = j + k

Next i

For L = 1 To M

LE = 2 ^ L

L1 = LE / 2

UR = 1#

UI = 0#

WR = Cos(Pi / L1)

WI = SN * Sin(Pi / L1)

For j = 1 To L1

For i = j To N Step LE

IP = i + L1

TR = AR(IP) * UR - AI(IP) * UI

TI = AI(IP) * UR + AR(IP) * UI

AR(IP) = AR(i) - TR

AI(IP) = AI(i) - TI

AR(i) = AR(i) + TR

AI(i) = AI(i) + TI

Next i

US = UR

UR = US * WR - UI * WI

UI = UI * WR + US * WI

Next j

Next L

If SN <>-1 Then

For i = 1 To N

AR(i) = AR(i) / N

AI(i) = AI(i) / N

Next i

End If

End Function

Private Function NTOM(N As Integer) As Integer

Dim ND As Double

ND = N

NTOM = 0

While (ND >1)

ND = ND / 2

NTOM = NTOM + 1

Wend

End Function

'*使用**********

Const fftIn = 128

Dim i As Integer

Dim xr(128) As Double

Dim xi(128) As Double

'赋值，IaIn(i)是采得的数据。

For i = 0 To 128

xr(i) = 100 * IaIn(i)

xi(i) = 0

Next

'FFT变换

Call FFT0(xr(), xi(), 128, 1)

'绘图

picI_FFT.Scale (0, 100)-(fftIn - 1, -10)

picI_FFT.DrawWidth = 2

For i = 0 To fftIn - 1

picI_FFT.Line (i, Abs(xr(i)))-(i + 1, Abs(xr(i + 1))), vbBlue

Next i