Simple Classifier Design coding

  • Status Closed
  • Budget N/A
  • Total Bids 5

Project Description

A team of scientists has been studying patients who might have a dangerous disease.

Medication is available but has serious side-e ects, so it is important to screen patients before they are given

the drug. The only known test, unfortunately, is very expensive, but it is known that some of the genes

implicated in the disease are also involved in the regulation of blood pressure and blood sodium levels, and

there is hope that it may be possible to screen subjects based on these two easily measured quantities. A

database of sodium levels, (L) and blood pressure (P) has been compiled, with patients labeled as positive

(D=1) or negative (D=0) for the disease based on the expensive test. Your task is to design a simple classi fier

that can classify future subjects with unknown labels correctly.

You will try three di erent approaches to the task:

1. A minimum distance classi er.

2. An aumented k-nearest neighbors classi er.

3. A single perceptron.

The augmented k-NN classi er will use one or more heuristics, such as using a customized k value for each

point based on some criterion (e.g., at least 60% of neighbors in the winning class), or using distance weights.

The choice of augmentation is up to you. You can think of it yourself or search in the literature for options.

You should write programs that implement each classi er and run them on the data set. You should use the

available data in a way that validates the ability of the classi ers to classify novel

This can be done using any coding language.

Get free quotes for a project like this

Looking to make some money?

  • Set your budget and the timeframe
  • Outline your proposal
  • Get paid for your work

Hire Freelancers who also bid on this project

    • Forbes
    • The New York Times
    • Time
    • Wall Street Journal
    • Times Online