Professional Documents
Culture Documents
ISSN:2320-0790
Hochschule Darmstadt
Abstract: By an approximat ion with a wavelet base we have in practise not only an error if the function y is not in Vj .
There we have a second error because we do not use all bases functions. If the wavelet has a compact support we have no
error by using only a part of all basis function. If we need an approximation on a compact interval I (which we can do even
if y is not quadratic integrable on R, because in that case it must only be quadratic integrab le on I) leads to worse
approximations if we calculate an orthogonal projection from 1 I y in Vj . We can get much better approximat ions, if we apply
a least square approximat ion with points in I. Here we will see, that this approximation can be much better th an a orthogonal
projection form y or 1 I y in Vj . With the Shannon wavelet, which has no compact support, we saw in many simulat ions, that
a least square approximat ion can lead to much better results than with well known wavelets with co mpact support. So in that
article we do an error estimat ion for the Shannon wavelet, if we use not all bases coefficients.
Keywords: component; formatting; style; styling; (insert minimu m 5 to 8 key words)
I. INTRODUCTION
In the wavelet theory a scaling function is used, which
has properties that are defined in the MSA (mult i scale
analysis). Through the MSA we know, that we can
construct an orthonormal basis of a closed subspace Vj ,
where Vj belongs to a sequence of subspaces with the
following property:
c
k
j ,k
c
k
j ,k
|k|kmax
2
k
|k|kmax
(1a)
y j (t ) :
y( t )
j ,k
( t )dt
ck y , j ,k
j ,k (t )
k k min
(1b)
kmax and kmin depend on the approximation interval I =
[t 0 ,tend].
ck 1I y , j ,k
y( t )
j ,k
( t )dt
340
x. ck = 1I y, j ,k . Additional
COMPUSOFT, An international journal of advanced computer technology, 2 (11), November-2013 (Volume-II, Issue-XI)
(2)
c
k
j ,k
j ,k
|k | k max
2 j/2c
(2)
|k | k max
k
k 1
Q (c )
(3)
j (t i )
y (t i )
i 0
ck
y
m
k0
y(t )
j , k (t ) dt
y(t )
j , k (t ) dt
y y jR
tI
L2 ( I )
if min y g
: c
gV j
y yI j
L2 ( I )
y y jR
sin( (2 j t k ))
max j ,k (t ) 2 j / 2 max
tI
(2 j t k )
tI
2 j/2
max
tI
1
2jt k
min y g
gV j
L2 ( I )
y y jI
L2 ( I )
Because of k > 2j a
II. EXAMPLES
2jt k
max j ,k (t )
2 j/2
k 2 a
tI
1) y is in L2 (R):
and
ck
2 j/2c
k 2ja
a) Orthogonal projection on V1 :
Now we get:
Here is the graph of y (dashed) and yj :
c
k
j ,k
c
k
j ,k
2 j/2c
|k | k max
k k max 1 (k
1.0
2 a)
j
0.9
0.8
If we set k0 =
=
0.7
0.6
0.5
c
k
j ,k
c
k
|k | k max
j ,k
2 j/2c
k k 0 1
0.4
1.0
k2
Figure 1.
341
0.5
0.0
0.5
1.0
COMPUSOFT, An international journal of advanced computer technology, 2 (11), November-2013 (Volume-II, Issue-XI)
The graph of y - yj :
t = 1/20.
0.9
0.8
1.0
0.5
0.5
1.0
0.7
0.6
5. 10 6
0.5
0.4
Figure 2.
1.0
Figure 5.
1.0
0.5
0.0
0.5
1.0
The graph of y - yj :
1. 10 10
0.8
5. 10 11
0.6
1.0
0.5
0.5
1.0
5. 10 11
0.4
1. 10 10
0.2
1.5 10 10
2. 10 10
2.5 10 10
Figure 3.
Figure 6.
For y - yj on [-3,3]:
Here we see that the approximat ion on the interval [-1, 1]
can even be used for an ext rapolation:
5. 10 6
0.8
0.6
5. 10 6
0.4
0.2
Figure 4.
1
0.2
Figure 7.
COMPUSOFT, An international journal of advanced computer technology, 2 (11), November-2013 (Volume-II, Issue-XI)
15
0.35
14
0.30
13
12
0.25
11
10
0.20
9
0.15
8
7
6
5
4
3
2
1
0.10
0.05
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Figure 8.
ckL(R)
k
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
9.98646 107
1.12346 106
1.27115 106
1.44692 106
1.65714 106
1.90945 106
2.2138 106
2.49637 106
6.39081 106
0.0000837542
0.00136911
0.0129464
0.0745338
0.260124
0.550701
0.707101
0.550701
0.260124
0.0745338
0.0129464
0.00136911
0.0000837542
6.39081 106
2.49637 106
2.2138 106
1.90945 106
1.65714 106
1.44692 106
1.27115 106
1.12346 106
9.98646 107
ck
0.0549149
0.162391
0.156844
0.0958456
0.11707
0.143967
0.00448466
0.309462
0.0238892
0.227517
0.0770245
0.00674154
0.0744649
0.26013
0.550695
0.707107
0.550695
0.26013
0.0744648
0.00663831
0.0799417
0.247064
0.0756419
0.193832
0.254212
0.282422
0.125028
0.124006
0.0955985
0.0000702266
0.287446
54 988.4
144 544.
123 386.
66 242.3
70 647.
75 396.
2026.77
123 966.
3739.06
2717.49
57.2587
0.479275
0.000924883
0.0000222773
0.0000111669
8.87617 106
0.0000111669
0.0000222773
0.000926093
0.487248
59.3894
2950.86
11 837.
77 646.3
114 832.
147 907.
75 447.2
85 704.4
75 207.3
63.5093
287 835.
ck
ck
|(ck
L( R)
- ck)/ ck
0.0555138
0.00279287
0.0908575
0.0753896
0.182931
0.274183
0.0337608
0.354246
0.0793108
0.250574
0.0757508
0.00727581
0.0744744
0.26013
0.550695
0.707107
0.550695
0.26013
0.0744823
0.0086523
0.0482002
0.115958
0.0431607
0.0832123
0.226748
0.00943628
0.154941
0.186859
0.0550771
0.132146
0.138983
ckL(R):
8.9239210-6
least square ck with t=1/20:
9.0077110-11
least square ck with t=1/100:
6.412210-11
y yj
L2 ([ 1.5,1.5])
ckL(R):
0.0000107337
L(R)
L2 ([ 1,1])
55 588.1
2486.95
71 475.6
52 102.6
110 391.
143 592.
15 249.2
141 905.
12 411.1
2992.78
56.3283
0.438006
0.000797106
0.0000222771
0.0000111669
8.87617 106
0.0000111669
0.000022277
0.000690303
0.331684
36.2054
1385.51
6752.56
33 334.3
102 426.
4940.88
93 498.1
129 144.
43 327.5
117 623.
139 171.
approximation methods:
9.98646 107
1.12346 106
1.27115 106
1.44692 106
1.65714 106
1.90945 106
2.2138 106
2.49637 106
6.39081 106
0.0000837542
0.00136911
0.0129464
0.0745338
0.260124
0.550701
0.707101
0.550701
0.260124
0.0745338
0.0129464
0.00136911
0.0000837542
6.39081 106
2.49637 106
2.2138 106
1.90945 106
1.65714 106
1.44692 106
1.27115 106
1.12346 106
9.98646 107
COMPUSOFT, An international journal of advanced computer technology, 2 (11), November-2013 (Volume-II, Issue-XI)
0.0000248848
0.0000196081
2) y is in V1 :
0.5
0.5
25
y (t )
k 25
For y - yj on [-3,3]:
a) Orthogonal projection on V1 :
0.0003
0.0002
0.0001
1
0.0001
0.0002
0.0003
1.0
Figure 12.
0.5
0.5
0.5
1.0
1.0
Figure 9.
0.5
The graph of y - yj :
1.0
0.5
0.5
0.00008
0.00006
Figure 13.
0.00004
0.00002
1.0
0.5
0.5
0.5
1.0
0.00002
The graph of y - yj :
Figure 10.
344
1.0
COMPUSOFT, An international journal of advanced computer technology, 2 (11), November-2013 (Volume-II, Issue-XI)
10.
9.
8.
7.
6.
5.
4.
3.
2.
1.
4. 10 10
3. 10 10
2. 10 10
1. 10 10
1.0
0.5
0.5
1.0
0.
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
1. 10 10
2. 10 10
Figure 14.
0.00990099
0.0121951
0.0153846
0.02
0.027027
0.0384615
0.0588235
0.1
0.2
0.5
1.
0.5
0.2
0.1
0.0588235
0.0384615
0.027027
0.02
0.0153846
0.0121951
0.00990099
4.16328
5.46063
0.84032
0.911112
0.0585332
0.0363541
0.0678428
0.0998917
0.2
0.5
1.
0.5
0.2
0.0998525
0.0746509
0.184991
0.805369
0.396838
0.6829
0.626755
1.40537
419.491
448.772
53.6208
44.5556
3.16573
1.94521
0.153328
0.00108337
7.19026 1011
1.75861 1010
9.0701 1011
1.52477 1010
8.76749 1011
0.00147549
0.269064
5.80977
28.7987
20.8419
45.3885
52.3939
140.943
1.0
0.5
k
3
15.
14.
13.
12.
11.
10.
9.
8.
7.
6.
5.
4.
3.
2.
0.5
Figure 15.
1.
0.
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
0.05
0.05
0.10
0.15
0.20
0.25
Figure 16.
ckL(R)
ck
ckL(R)
0.00442478
0.00507614
0.00588235
0.00689655
0.00819672
0.00990099
0.0121951
0.0153846
0.02
0.027027
0.0384615
0.0588235
0.1
0.2
0.5
1.
0.5
0.2
0.1
0.0588235
0.0384615
0.027027
0.02
0.0153846
0.0121951
0.00990099
0.00819672
0.00689655
0.00588235
0.00507614
0.00442478
ck
0.171272
0.0487982
0.497604
0.247044
0.117465
0.621908
0.0267993
0.196127
0.205125
0.177672
0.1073
0.0531263
0.10006
0.2
0.5
1.
0.5
0.2
0.100071
0.0515767
0.135855
0.308153
0.318697
0.0523028
0.204147
0.0272149
0.137145
0.0496725
0.218437
0.209121
0.0107344
39.7075
10.6132
85.5926
34.8214
15.3307
61.8127
1.19754
13.7482
9.25624
7.57388
1.78981
0.096853
0.000604441
6.2635 1010
8.38828 1011
4.66002 1011
8.86247 1011
6.92697 1010
0.000711496
0.123196
2.53223
12.4017
14.9348
4.39968
15.74
3.74871
15.7316
6.20252
38.1342
42.1968
3.42598
COMPUSOFT, An international journal of advanced computer technology, 2 (11), November-2013 (Volume-II, Issue-XI)
y yj
2. 10 10
L2 ([ 1,1 ])
approximation methods:
1. 10 10
5. 10 11
ck
L( R)
:
1.0
0.5
0.5
0.0000708509
5. 10 11
1. 10 10
1.0
1.5 10 10
1.8087510-10
Figure 17.
0.25
y yj
L2 ([ 1.5,1.5])
0.20
0.15
ckL(R):
0.10
0.00013277
0.05
0.000050488
Figure 18.
The graph of y - yj :
346
COMPUSOFT, An international journal of advanced computer technology, 2 (11), November-2013 (Volume-II, Issue-XI)
Figure 21.
2.5
2.0
1.5
The graph of y - yj :
1.0
0.5
1. 10 9
1.0
0.5
0.5
1.0
1.0
Figure 19.
0.5
0.5
1.0
1. 10 9
2. 10 9
The graph of y - yj :
3. 10 9
0.3
Figure 22.
0.2
0.1
0.5
0.5
1.0
0.1
20
0.2
15
Figure 20.
10
Figure 23.
2.5
2.0
1.5
1.0
0.5
1.0
0.5
0.5
1.0
347
COMPUSOFT, An international journal of advanced computer technology, 2 (11), November-2013 (Volume-II, Issue-XI)
3.0
2.5
4.9004510-10
2.0
1.5
y yj
1.0
L2 ([ 1.5,1.5])
0.5
Figure 24.
ckL( R)
k
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
0.045956
0.0505131
0.0560689
0.062991
0.0718511
0.0835905
0.0998736
0.123941
0.163037
0.237215
0.428941
2.34877
3.6603
1.6616
1.34176
0.574112
0.536393
0.168756
0.240166
0.0124894
0.0488908
0.0473013
0.0440978
0.0409305
0.0380603
0.0355111
0.0332544
0.0312522
0.0294685
0.0278719
0.0264356
ck
0.636832
3.9634
0.40329
1.79733
0.709774
1.46529
2.04353
1.32979
5.97515
1.31779
6.26748
5.09408
3.16799
1.92212
1.16582
0.707107
0.428882
0.26013
0.157235
0.0470357
0.492759
1.32103
0.00904417
0.563427
0.99555
1.25702
0.215155
0.117797
0.369592
0.961601
1.18362
L2 ([ 1,1])
ckL(R):
14.8574
77.4629
8.19276
29.5331
10.8784
18.5293
21.4612
11.7292
35.649
4.55524
15.6115
1.16883
0.134501
0.156783
0.131127
0.231653
0.200433
0.541458
0.345307
4.76604
11.0788
26.928
1.20509
14.7655
27.1572
36.3981
5.46998
2.76924
11.5419
33.5008
43.7738
0.353685
least square ck with t=1/ 20:
0.00032093
least square ck with t=1/ 100:
0.000178703
III. CONCLUSION
In all examp les the least square method (which is a
numerical approach to get the "best" solution on the
approximation interval I) leads to very good results. Even
to better results as with the best L2 approximation on R.
approximation methods:
ckL( R):
0.219725
COMPUSOFT, An international journal of advanced computer technology, 2 (11), November-2013 (Volume-II, Issue-XI)
F. Table Captions
Tables must be numbered using uppercase Roman
numerals. Table captions must be centred and in 8 pt
Regular font with Small Caps. Every word in a table
caption must be capitalized except for short minor words as
listed in Section III-B. Captions with table numbers must
be placed before their associated tables, as shown in Table
Sr.
No
.
Heading
1
Heading
2
Heading
3
Headin
g4
Heading
5
Headin
g6
IV. REFERENCES
The heading of the References section must not be
numbered. All reference items must be in 8 pt font. Please
use Regular and Italic styles to distinguish different fields
as shown in the References section. Nu mber the reference
items consecutively in square brackets (e.g. [1]).
E. Figure Captions
349