Do the following, integrate the results in a Word document. The document should be well formatted mixing the graphs and explanations as requested in the assignment, with ABSOLUTELY NO handwritten or hand drawn items.
Matlab Issues
Here is an example from Matlab's help on how to generate a 5-by-5 array of random numbers with a mean of .6 that are distributed with a variance of 0.1:
x = .6 + sqrt(0.1) * randn(5)
x =
0.8713 0.4735 0.8114 0.0927 0.7672
0.9966 0.8182 0.9766 0.6814 0.6694
0.0960 0.8579 0.2197 0.2659 0.3085
0.1443 0.8251 0.5937 1.0475 -0.0864
0.7806 1.0080 0.5504 0.3454 0.5813
Matlab Assignment
1) Do problem 3.4 from the textbook. Use normalized input vectors to create the weight matrix. Test network performance when the input vectors are corrupted with zero-mean Gaussian noise (not white noise). Vary the power of the noise by changing the standard deviation from 0 to 1 in steps of 0.1. Note that when the standard deviation is set to zero you are testing the performance of the network with noiseless inputs. Show a graph of the probability of misclassification for the 11 cases of noise corruption. To estimate these probabilities apply all the input vectors to the network one at a time and count the number of misclassifications. To obtain a reliable estimate repeat this test 400 times for each value of noise power. What is the probability of misclassification when the standard deviation is 0? What does this result mean? Repeat for part b) for all 11 values of standard deviation (not only for standard deviation 0.3) and show a new graph of probabilities.