1. The technical description of all techniques utilized, 2. The design of the algorithms (pseudo-code, flowcharts, or some other structured descriptive means), 3. The results of the algorithms, 4. An analysis of the results, i.e., did you obtain what you expected? Were there any surprises? What conclusions can you draw from the experiments? etc.
For this experiment you are to implement a multi-layer perceptron containing a single hidden layer with 10 neurons. Download the cross dataset from Canvas. The zip file contains the dataset and initial weights you should use for this experiment. The file cross_data.csv contains 314 two-dimensional samples, each with a target value of 0 or 1 The initial weights and biases for this network are listed in the tables below, and are also given in the files w1.csv, bl.csv, w2.csv, and b2.csv. b2.csv Node Bias bi0.1131 bl.csv w1.csvFrom Input Node (weights) NodeBias b1(1) 1-0-122 b0.9401 b 0.4271 b4) b) bo) 0.3326 b 0.6961 b -0.9316 bo 0.3681 bio) 1.0695 W2) 0.39 0.6544 0.6376 -0.0601 Wa(1) | 0.0064 | -0.0462 0.0782 0.2728 -0.2115 1.0252 -0.1775 -0.7019 ws wo0 wg) -0.7109 0.349 wio0.8441 0.4276 w2.csv From Hidden Node 0 Output w0.0511 0.1611 0.0238 -0.0267 0.1089 0.2381 0.0784 0.003 0.1646 -0.1779 Node Use the following parameters in your network Sigmoid activation function for all nodes: (v) . L+expz) Learning rate: Momentum term: = 0.7 =0.3 1) Perform one epoch of on-line backpropagation training (update weights after each sample) on the cross dataset in the order provided. Do not randomize the sample presentation order for the first epoch. List all of the network weight and bias terms after the first epoch using For this experiment you are to implement a multi-layer perceptron containing a single hidden layer with 10 neurons. Download the cross dataset from Canvas. The zip file contains the dataset and initial weights you should use for this experiment. The file cross_data.csv contains 314 two-dimensional samples, each with a target value of 0 or 1 The initial weights and biases for this network are listed in the tables below, and are also given in the files w1.csv, bl.csv, w2.csv, and b2.csv. b2.csv Node Bias bi0.1131 bl.csv w1.csvFrom Input Node (weights) NodeBias b1(1) 1-0-122 b0.9401 b 0.4271 b4) b) bo) 0.3326 b 0.6961 b -0.9316 bo 0.3681 bio) 1.0695 W2) 0.39 0.6544 0.6376 -0.0601 Wa(1) | 0.0064 | -0.0462 0.0782 0.2728 -0.2115 1.0252 -0.1775 -0.7019 ws wo0 wg) -0.7109 0.349 wio0.8441 0.4276 w2.csv From Hidden Node 0 Output w0.0511 0.1611 0.0238 -0.0267 0.1089 0.2381 0.0784 0.003 0.1646 -0.1779 Node Use the following parameters in your network Sigmoid activation function for all nodes: (v) . L+expz) Learning rate: Momentum term: = 0.7 =0.3 1) Perform one epoch of on-line backpropagation training (update weights after each sample) on the cross dataset in the order provided. Do not randomize the sample presentation order for the first epoch. List all of the network weight and bias terms after the first epoch using
