Activation Functions in Neuron Perceptron Networks Assignment

Assignment Task

Questions

1) Do the following tasks on the single neuron perceptron program discussed in Lecture 4.

• Update the program for classification along the line ax + by + c = 0
• Use sign() as activation function
•  Use sigmoid() as activation function
• How do you preprocess the data points and targets to apply sigmoid()?
• Do they produce the same classification line for the same data points and initial weights? Justify.
• Update the program for 3D classification plane ax + by + c
• Randomly generate 1000 points in the cuboid 700 × 700 × 500 and classify points with a plane of your choice.
• Apply above all 4 subparts. Take a sample from 1000 points and train using both activation functions and test with remaining points for classification.
• What least possible sample size is enough to train and test all points to classify correctly.

Use the data given in Lecture-2(Linear Regression) for 25 house sizes and prices.

• Write a Neural Network program to fit the line.
• Compare the result with the linear regression line and comment on your findings. 

This IT and Computer Science has been solved by our PHD Experts at My Uni Paper.