1. Implementing fisher LDA method using MATLAB.
- Write functions that take a data set and compute the optimal projection vector following the fisher criterion.
- The input set of instances can be of two dimensions or of more than two dimensions.
- The output from the main function must be the identified projection vector.
2. Implementing perceptron method using MATLAB.
- Write functions that take a set of examples and identify the hyper plane to separate the examples
- The input set of instances can be of two dimensions or of more than two dimensions.
- The output from the main function must be the identified hyper plane.
3. Write functions to generate synthetic data for evaluating the LDA and perceptron methods using MATLAB.
- The functions should be able to allow user to specify how many dimensions of each instance, how many classes, and how many examples in each class.
- The functions should generate random instances and provide a label for each instance.
- Use the examples generated from the functions to evaluate the two methods developed in 1 and 2. Conduct cross-validation and report the average accuracy, sensitivity, and specificity.
NOTE:
a. There is no requirement to develop a GUI to visualize the intermediate or the final results, although a visualization is recommended.
b. The synthetic data generated using the functions from problem 3 (above) are suggested to be used for evaluating the two methods from problems 1 and 2.
A report that includes the following items is due in addition to the source code for the functions:
- A description of each function,
- How to run the functions to get the reported results, and
- The experimental results.