K-Nearest Neighbor (KNN) Predicts the label of a data point based on the K nearest data points; a higher K value averages over larger range of points in order to make an identifying prediction Results of the KNN Classifier on our Training Data: |
http://bdewilde.github.io/blog/blogger/2012/10/26/classification-of-hand-written-digits-3/
|
This scatter plot displays correctly identified points as dots and incorrectly identified points with an 'x'
- Had the highest accuracy of 94.8% |
This confusion matrix plots predicted label on the x axis and actual label on the y axis. If they are the same, the intersecting box is green. If then are different, the intersecting box is red.
|
Support Vector Machine (SVM)
Predicts the label of a data point based on which general group of data it appears to falls in; creates sections in the data by drawing lines in gaps in the data; the separating lines can be linear, quadratic, cubic, etc. depending on how specified. In this case, they are cubic functions Results of the SVM Classifier on our Training Data: |
https://en.wikipedia.org/wiki/Support_vector_machine#/media/File:Kernel_Machine.svg
|