Crab Classification
This demo illustrates using a neural network as a classifier to identify the sex of crabs from physical dimensions of the crab.
Contents
The Problem: Classification of Crabs
In this demo we attempt to build a classifier that can identify the sex of a crab from its physical measurements. Six physical characterstics of a crab are considered: species, frontallip, rearwidth, length, width and depth. The problem on hand is to identify the sex of a crab given the observed values for each of these 6 physical characterstics.
Why Neural Networks?
Neural networks have proven themselves as proficient classifiers and are particularly well suited for addressing non-linear problems. Given the non-linear nature of real world phenomena, like crab classification, neural networks is certainly a good candidate for solving the problem.
The six physical characterstics will act as inputs to a neural network and the sex of the crab will be target. Given an input, which constitutes the six observed values for the physical characterstics of a crab, the neural network is expected to identify if the crab is male or female.
This is achieved by presenting previously recorded inputs to a neural network and then tuning it to produce the desired target outputs. This process is called neural network training.
Preparing the Data
Data for classification problems can very often have textual or non-numeric information. In our case, sex of the crab is non-numeric (Male/Female). Neural networks however cannot be trained with non-numeric data. Hence we need to translate the textual data into a numeric form.
There are several ways to translate textual or symbolic data into numeric data. Some of the common symbol translation techniques used are unary encoding, binary encoding and numbering classes. We are going to use unary encoding in this demo to perform symbol translation.
fid = fopen('private/crabdata.csv'); C = textscan(fid,'%f%f%f%f%f%f%s','delimiter',','); % Import data fclose(fid);
The first 6 columns of data represent the crab's physical characterstics. The 7th column represents the sex of the crab.
physchars = [C{1} C{2} C{3} C{4} C{5} C{6}]; % inputs to neural network
female = strncmpi(C{7}, 'Female', 1);
male = strncmpi(C{7}, 'Male', 1);
sex = double([female male]); % targets for neural network
'Female' is now represented by the vector [1 0] and 'Male' is represented by the vector [0 1]. This process is called unary encoding. The neural network will now be trained to produce a [1 0] output for 'Female' crabs and [0 1] output for 'Male' crabs.
We could also have used numbers to represent the two sexes (Male=1, Female=2) or performed binary encoding (Male=[0 0], Female=[0 1]).
The next step is to preprocess the data into a form that can be used with a neural network.
The neural network object in the toolbox expects the samples along columns and its features along rows. Our dataset has its samples along rows and its features along columns. Hence the matrices have to be transposed.
physchars = physchars'; sex = sex';
Building the Neural Network Classifier
The next step is to create a neural network that will learn to identify the sex of the crabs.
Since the neural network starts with random initial weights, the results of this demo will differ slightly every time it is run. The random seed is set to avoid this randomness. However this is not necessary for your own applications.
rand('seed', 491218382)
A 1-hidden layer feed forward network is created with 20 neurons in the hidden layer.
net = newff(physchars,sex,20); % Create a new feed forward network
Now the network is ready to be trained. The samples are automatically divided into training, validation and test sets. The training set is used to teach the network. Training continues as long as the network continues improving on the validation set. The test set provides a completely independent measure of network accuracy.
[net,tr] = train(net,physchars,sex);
Testing the Classifier
The trained neural network can now be tested with the testing samples This will give us a sense of how well the network will do when applied to data from the real world.
testInputs = physchars(:,tr.testInd);
testTargets = sex(:,tr.testInd);
out = sim(net,testInputs); % Get response from trained network
The network response can now be compared against the desired target response to build the classification matrix which will provides a comprehensive picture of a classifiers performance.
[y_out,I_out] = max(out); [y_t,I_t] = max(testTargets); diff = [I_t - 2*I_out]; f_f = length(find(diff==-1)); % Female crabs classified as Female f_m = length(find(diff==0)); % Female crabs classified as Male m_m = length(find(diff==-2)); % Male crabs classified as Male m_f = length(find(diff==-3)); % Male crabs classified as Female N = size(testInputs,2); % Number of testing samples fprintf('Total testing samples: %d\n', N); cm = [f_f f_m; m_f m_m] % classification matrix
Total testing samples: 40
cm =
18 0
1 21
The classification matrix provides a comprehensive picture of the classification performance of the classifier. The ideal classification matrix is the one in which the sum of the diagonal is equal to the number of samples.
It can also be understood in terms of percentages. The following matrix provides the same information as above but in terms of percentages.
cm_p = (cm ./ N) .* 100 % classification matrix in percentages fprintf('Percentage Correct classification : %f%%\n', 100*(cm(1,1)+cm(2,2))/N); fprintf('Percentage Incorrect classification : %f%%\n', 100*(cm(1,2)+cm(2,1))/N);
cm_p =
45.0000 0
2.5000 52.5000
Percentage Correct classification : 97.500000%
Percentage Incorrect classification : 2.500000%
This demo illustrated using a neural network to classify crabs.
Explore other demos and the documentation for more insight into neural networks and its applications.