k-Nearest neighbors in R with the 'class' package

preview_player
Показать описание
You can download the source code at this link:
Рекомендации по теме
Комментарии
Автор

Man got mad skills in R. Great stuff, big man. Thanks for the help

saltyYagi_
Автор

Thanks for this info. it was to the point and helped me out using the 'class' library!

jasonjiradev
Автор

Great stuff... code on git hub was also very useful. Some other nice nuggets besides the actual knn- a neat way to sample data, and a function for clean output... much thanks!

baruchschwartz
Автор

What if you want to apply the algorithm to factors rather than integers... for example cities, or other classifications... how do you convert the factor into something that the computer can use to find euclidean distance?

baruchschwartz
Автор

How do we make the classification of a completely new dataset baased on this model?

tansutazegul
Автор

Hi. Is KNN only work for numeric variables? If one of my variable is a categorical variables (e.g. High / Mid/ Low), can I still use KNN?

SC-bimy
Автор

Thanks, Dr. Castro. You are very adept with R coding. I hope to get there one day. Is there a simple way to choose n to minimize false negative while maintaining reasonable (e.g. 0.8) accuracy?

garykuleck
Автор

Thanks for the tutorial. Can we perform KNN in steam player data set? It has 6 columns.[Game Name, Month, Avg Players per month, Peak Player per month, avg gain, avg % gain] I am planning to implement knn for players count for the games. Based on the number of players, I want to classify the games in groups

SmartphoneWorldTamil
Автор

What is the data actually about? What do the X1, X2 values represent?

joannathornett