![ART1 Example
Step1. Initialize Parameters
Input Units (n) = 9,
Cluster Units (m) = 2, L = 2
Input Pattern is [ 1 1 1 1 0 1 1 1 1]
Step2. While stopping condition is not satisfied,
do Steps 3-14
Step3. For each training input pattern, do Steps 4-13
Here training input pattern is [ 1 1 1 1 0 1 1 1 1]
4.0
5
1
2
1
5
1
0
5
1
2
1
5
1
0
5
1
2
1
5
1
0
5
1
2
1
5
1
0
5
1
2
1
ijb
111111111
101010101
jit](https://image.slidesharecdn.com/ann-3-200426112852/85/Ann-3-6-320.jpg)
![ART1
Step 4. Set activation of all F2 units is zero
Set activation of all F1(a) units is input vector
s = [ 1 1 1 1 0 1 1 1 1]
Step 5. Compute norm of s
Step.6 Send input signal from F1(a) to F1(b) layer
xi = si
x = [ 1 1 1 1 0 1 1 1 1]
Step.7 Calculate net input
y1 = 2
y2 = 1.6
y1 > y2
J = 1
8
1
n
i
iss
n
i
ijij bxy
1](https://image.slidesharecdn.com/ann-3-200426112852/85/Ann-3-7-320.jpg)
![ART1
Step 8. While reset is true do Steps 9-12
Step 9. The winning unit is
J = 1
Step 10. Recompute activation of F1
x = [ 1 1 1 1 0 1 1 1 1] [ 1 0 1 0 1 0 1 0 1]
x = [ 1 0 1 0 0 0 1 0 1]
Step 11. Calculate norm of x
Step 12. Test for reset
Go to Step 13
Else
yJ = -1, proceed to step 8
Jiii tsx
4
1
n
i
ixx
s
x
s
x
,4.05.08/4
falseisresetthen
s
x
if ](https://image.slidesharecdn.com/ann-3-200426112852/85/Ann-3-8-320.jpg)
Ann 3
- 1. Soft Computing: Artificial Neural Networks-Unsupervised Model-ART1 Dr. Baljit Singh Khehra Professor CSE Department Baba Banda Singh Bahadur Engineering College Fatehgarh Sahib-140407, Punjab, India
- 2. ART1 Binary Adaptive Resonance Theory Designed to perform for binary input vector Unsupervised Learning ART1 has F1 units (Input and Interface units) F2 units (Cluster units) A Reset unit: implements user Control over the degree of similarity of patterns placed on the same cluster
- 3. ART1 Step1. Initialize Parameters Input Units (n) , Cluster Units (m) , L >1 Vigilance parameter ( ) Bottom-up weights (bij) Top-down weights Tji (0) =1 Step2. While stopping condition is not satisfied, do Steps 3-14 Step3. For each training input pattern, do Steps 4-13 Here training input pattern is s Step 4. Set activation of all F2 units is zero Set activation of all F1(a) units is input vector (s) 10 nL L bij 1 )0(0
- 4. ART1 Step 5. Compute norm of s Step.6 Send input signal from F1(a) to F1(b) layer xi = si Step.7 Calculate net input j=1,2,3,………,m i=1,2,3,……….,n Find yJ such that yJ ≥ yj for j=1,2,3,………,m Step 8. While reset is true do Steps 9-12 Step 9. The winning unit is J Step 10. Recompute activation of F1 n i iss 1 n i ijij bxy 1 Jiii tsx
- 5. ART1 Step 11. Calculate norm of x Step 12. Test for reset yJ = -1, proceed to step 8 Go to Step 13 Step13. Update weights Step14. Test for stopping condition n i ixx 1 Falseisresetthen s x if Trueisresetthen s x if xL Lx newb i iJ 1 )( iJi xnewt )(
- 6. ART1 Example Step1. Initialize Parameters Input Units (n) = 9, Cluster Units (m) = 2, L = 2 Input Pattern is [ 1 1 1 1 0 1 1 1 1] Step2. While stopping condition is not satisfied, do Steps 3-14 Step3. For each training input pattern, do Steps 4-13 Here training input pattern is [ 1 1 1 1 0 1 1 1 1] 4.0 5 1 2 1 5 1 0 5 1 2 1 5 1 0 5 1 2 1 5 1 0 5 1 2 1 5 1 0 5 1 2 1 ijb 111111111 101010101 jit
- 7. ART1 Step 4. Set activation of all F2 units is zero Set activation of all F1(a) units is input vector s = [ 1 1 1 1 0 1 1 1 1] Step 5. Compute norm of s Step.6 Send input signal from F1(a) to F1(b) layer xi = si x = [ 1 1 1 1 0 1 1 1 1] Step.7 Calculate net input y1 = 2 y2 = 1.6 y1 > y2 J = 1 8 1 n i iss n i ijij bxy 1
- 8. ART1 Step 8. While reset is true do Steps 9-12 Step 9. The winning unit is J = 1 Step 10. Recompute activation of F1 x = [ 1 1 1 1 0 1 1 1 1] [ 1 0 1 0 1 0 1 0 1] x = [ 1 0 1 0 0 0 1 0 1] Step 11. Calculate norm of x Step 12. Test for reset Go to Step 13 Else yJ = -1, proceed to step 8 Jiii tsx 4 1 n i ixx s x s x ,4.05.08/4 falseisresetthen s x if
- 9. ART1 Step13. Update weights Step14. Test for stopping condition 5 2 1 )( ii iJ x xL Lx newb iJi xnewt )( 5 1 5 2 5 1 0 5 1 5 2 5 1 0 5 1 0 5 1 0 5 1 5 2 5 1 0 5 1 5 2 )(newbij 111111111 101000101 )(newt ji