roiabr/machine-learning-knn

Implementation of K Nearest Neighbors in machine learing course

KNN Algorithm

About the project:

In this project I implemented the KNN algorithm in the machine learning course. The dataset is the Hope College Temperature data set. The data described by body temperature in degrees Fahrenheit, the gender (1 = male, 2 = female) and the heart rate in beats per minute.

We are running the knn algorithm 500 times for each of k = 1, 3, 5, 7, 9 when k is the number of neighbours. For each run, randomly dividing the points into 50% base points and 50% test points. Then run knn on p when p is the minkowski distance(1, 2, infinity), and after computing the final best neighbours, find its error T on every points in test and traing. Dataset contains 130 data points. The label (1 and -1) will be the gender, and the temperature and heartrate define the 2-dimensional point.

Roi Abramovitch