Machine Learning Algorithms: From Math to Code
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
Machine Learning Algorithms: From Math to Code
Assignment for BackPropagation
1 BackPropagation
For this problem, you should implement missing code in backPropagation.m to fulfill back propagation and write a report.
1.1 backPropagation.m
After you implement missing code in backPropagation.m and apply different learning rates, you should get the outputs in the command line as follows (taking eta as 1 for example). Elapsed time depends on the machine so pay no attention to it if it differs from the value provided, so as training process. Try different learning rates to see how learning rate affects the training process including training time, convergence and error rate and report them. Analyse how learning rate effects the training process.
>> backPropagation
d=3 N=400 K=8
Training:
epoch 100: er=0.0167
epoch 200: er=0.0113
epoch 300: er=0.0090
epoch 400: er=0.0076
epoch 500: er=0.0068
epoch 600: er=0.0061
epoch 700: er=0.0056
epoch 800: er=0.0052
epoch 900: er=0.0049
epoch 1000: er=0.0046
epoch 1100: er=0.0044
epoch 1200: er=0.0042
epoch 1300: er=0.0040
epoch 1400: er=0.0039
epoch 1500: er=0.0037
epoch 1600: er=0.0036
epoch 1700: er=0.0035
epoch 1800: er=0.0034
epoch 1900: er=0.0033
epoch 2000: er=0.0032
epoch 2100: er=0.0031
epoch 2200: er=0.0030
epoch 2300: er=0.0029
epoch 2400: er=0.0029
epoch 2500: er=0.0028
epoch 2600: er=0.0028
epoch 2700: er=0.0027
epoch 2800: er=0.0027
epoch 2900: er=0.0026
epoch 3000: er=0.0026
epoch 3100: er=0.0025
epoch 3200: er=0.0025
epoch 3300: er=0.0024
epoch 3400: er=0.0024
epoch 3500: er=0.0024
epoch 3600: er=0.0023
epoch 3700: er=0.0023
epoch 3800: er=0.0023
epoch 3900: er=0.0022
epoch 4000: er=0.0022
epoch 4100: er=0.0022
epoch 4200: er=0.0021
epoch 4300: er=0.0021
epoch 4400: er=0.0021
epoch 4500: er=0.0021
epoch 4600: er=0.0020
epoch 4700: er=0.0020
epoch 4800: er=0.0020
epoch 4900: er=0.0020
epoch 5000: er=0.0019
epoch 5100: er=0.0019
epoch 5200: er=0.0019
epoch 5300: er=0.0019
epoch 5400: er=0.0019
epoch 5500: er=0.0019
epoch 5600: er=0.0018
epoch 5700: er=0.0018
epoch 5800: er=0.0018
epoch 5900: er=0.0018
epoch 6000: er=0.0018
epoch 6100: er=0.0018
epoch 6200: er=0.0017
epoch 6300: er=0.0017
epoch 6400: er=0.0017
epoch 6500: er=0.0017
epoch 6600: er=0.0017
epoch 6700: er=0.0017
epoch 6800: er=0.0017
epoch 6900: er=0.0016
epoch 7000: er=0.0016
epoch 7100: er=0.0016
epoch 7200: er=0.0016
epoch 7300: er=0.0016
epoch 7400: er=0.0016
epoch 7500: er=0.0016
epoch 7600: er=0.0016
epoch 7700: er=0.0016
epoch 7800: er=0.0015
epoch 7900: er=0.0015
epoch 8000: er=0.0015
epoch 8100: er=0.0015
epoch 8200: er=0.0015
epoch 8300: er=0.0015
epoch 8400: er=0.0015
epoch 8500: er=0.0015
epoch 8600: er=0.0015
epoch 8700: er=0.0015
epoch 8800: er=0.0014
epoch 8900: er=0.0014
epoch 9000: er=0.0014
epoch 9100: er=0.0014
epoch 9200: er=0.0014
epoch 9300: er=0.0014
epoch 9400: er=0.0014
epoch 9500: er=0.0014
epoch 9600: er=0.0014
epoch 9700: er=0.0014
epoch 9800: er=0.0014
epoch 9900: er=0.0014
epoch 10000: er=0.0014
epoch 10100: er=0.0013
epoch 10200: er=0.0013
epoch 10300: er=0.0013
epoch 10400: er=0.0013
epoch 10500: er=0.0013
epoch 10600: er=0.0013
epoch 10700: er=0.0013
epoch 10800: er=0.0013
epoch 10900: er=0.0013
epoch 11000: er=0.0013
epoch 11100: er=0.0013
epoch 11200: er=0.0013
epoch 11300: er=0.0013
epoch 11400: er=0.0013
epoch 11500: er=0.0013
epoch 11600: er=0.0012
epoch 11700: er=0.0012
epoch 11800: er=0.0012
epoch 11900: er=0.0012
epoch 12000: er=0.0012
epoch 12100: er=0.0012
epoch 12200: er=0.0012
epoch 12300: er=0.0012
epoch 12400: er=0.0012
epoch 12500: er=0.0012
epoch 12600: er=0.0012
epoch 12700: er=0.0012
epoch 12800: er=0.0012
epoch 12900: er=0.0012
epoch 13000: er=0.0012
epoch 13100: er=0.0012
epoch 13200: er=0.0012
epoch 13300: er=0.0012
epoch 13400: er=0.0012
epoch 13500: er=0.0012
epoch 13600: er=0.0011
epoch 13700: er=0.0011
epoch 13800: er=0.0011
epoch 13900: er=0.0011
epoch 14000: er=0.0011
epoch 14100: er=0.0011
epoch 14200: er=0.0011
epoch 14300: er=0.0011
epoch 14400: er=0.0011
epoch 14500: er=0.0011
epoch 14600: er=0.0011
epoch 14700: er=0.0011
epoch 14800: er=0.0011
epoch 14900: er=0.0011
epoch 15000: er=0.0011
epoch 15100: er=0.0011
epoch 15200: er=0.0011
epoch 15300: er=0.0011
epoch 15400: er=0.0011
epoch 15500: er=0.0011
epoch 15600: er=0.0011
epoch 15700: er=0.0011
epoch 15800: er=0.0011
epoch 15900: er=0.0011
epoch 16000: er=0.0011
epoch 16100: er=0.0011
epoch 16200: er=0.0010
epoch 16300: er=0.0010
epoch 16400: er=0.0010
epoch 16500: er=0.0010
epoch 16600: er=0.0010
epoch 16700: er=0.0010
epoch 16800: er=0.0010
epoch 16900: er=0.0010
epoch 17000: er=0.0010
epoch 17100: er=0.0010
epoch 17200: er=0.0010
epoch 17300: er=0.0010
epoch 17400: er=0.0010
epoch 17500: er=0.0010
epoch 17600: er=0.0010
Converged in 100000 iterations: true
0/200
Elapsed time 60.116041 seconds.
Testing:
|
0/200 Cm = |
|
|
|
|
|
|
|
|
27 |
0 |
0 |
0 |
0 |
0 |
0 |
0 |
|
0 |
27 |
0 |
0 |
0 |
0 |
0 |
0 |
|
0 |
0 |
27 |
0 |
0 |
0 |
0 |
0 |
|
0 |
0 |
0 |
22 |
0 |
0 |
0 |
0 |
|
0 |
0 |
0 |
0 |
21 |
0 |
0 |
0 |
|
0 |
0 |
0 |
0 |
0 |
24 |
0 |
0 |
|
0 |
0 |
0 |
0 |
0 |
0 |
27 |
0 |
|
0 |
0 |
0 |
0 |
0 |
0 |
0 |
25 |
Testing: er=0.0000
Notes
1. These two problems should be included in a single report with headings.
2. Source code and report should be compressed into a single .zip file and handed on the canvas before next monday midnight, July 24 23:59.
2023-07-22
Assignment for BackPropagation