Backpropagation Neural Networks (Python)

Introduction

This assignment is to investigate back-propagation neural networks as applied

to classifying flowers. The requirements are as follows:

1.Code the back-propagation algorithm presented in class.

2.Identify the proper network topology to solve the flower recognition problem.

3.Train and test the network over the flowers data set.

4.Capture prediction accuracies (total number of correct predictions over the

testing examples) of the network.

5.If unacceptable performance, then return to Step 2.

6.Write up your results.

Problem

The objective is to classify color images of flowers into one of 11 flower types.

Programming

Construct, train, and test the most accurate back-propagation neural network

you can to classify the flower data. The following are guidelines to construct and

train your neural net:

Code: You must develop your own code based on the algorithm presented in

our lecture slides.

Topology of your Neural Net: Your network should consist of 1 input layer, 1

hidden layer (N neurons), and 1 output layer (11 neurons).

Training and Testing: Randomly subdivided the data set into 70% training and

30% testing images. Prediction accuracy will be based on the average error

observed over the 30% of testing images.

Data

flower_images.zip in the files directory