MINND
Package
weka.classifiers.mi
Synopsis
Multiple-Instance Nearest Neighbour with Distribution learner.
It uses gradient descent to find the weight for each dimension of each exeamplar from the starting point of 1.0. In order to avoid overfitting, it uses mean-square function (i.e. the Euclidean distance) to search for the weights.
It then uses the weights to cleanse the training data. After that it searches for the weights again from the starting points of the weights searched before.
Finally it uses the most updated weights to cleanse the test exemplar and then finds the nearest neighbour of the test exemplar using partly-weighted Kullback distance. But the variances in the Kullback distance are the ones before cleansing.
For more information see:
Xin Xu (2001). A nearest distribution approach to multiple-instance learning. Hamilton, NZ.
Options
The table below describes the options available for MINND.
Option |
Description |
|---|---|
debug |
If set to true, classifier may output additional info to the console. |
numNeighbours |
The number of nearest neighbours to the estimate the class prediction of test bags. |
numTestingNoises |
The number of nearest neighbour instances in the selection of noises in the test data. |
numTrainingNoises |
The number of nearest neighbour instances in the selection of noises in the training data. |
Capabilities
The table below describes the capabilites of MINND.
Capability |
Supported |
|---|---|
Class |
Missing class values, Binary class, Nominal class |
Attributes |
Unary attributes, Binary attributes, Empty nominal attributes, Relational attributes, Missing values, Nominal attributes |
Other |
Only multi-Instance data |
Min # of instances |
1 |