KNN only predicting one class. Using python and sklearn

Question

I have a program that works flawlessly on known data sets such as the Breast Cancer Wisconsin data. I have created a novel data set with current air pressure in 100 cities and the future movement of a stock price. So about 100 dimensions plus the classifier which is -1 or 1. There are about 350 instances of data. The algorithm is showing about an 80% accuracy.

My problem is that when I input old data to make a prediction, it is only predicting 1, never -1. This is even the case when the training data contained the -1 classifier. Any ideas are greatly appreciated!

simon · Accepted Answer · 2016-12-07 11:26:00Z

2

KNN takes a vote on the n nearest neighbours. So with n=5 you need 3 neighbours to vote -1. This may be never if -1 is a rare occurrence and the -1s are not clustered together. Suggest you try other algorithms as well.

answered Dec 7, 2016 at 11:26

simon

2,9112 gold badges21 silver badges30 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Hanley Soilsmith Over a year ago

Thanks, that makes sense :)

Collectives™ on Stack Overflow

KNN only predicting one class. Using python and sklearn

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related