Scribd
Upload
76 views

MLT Notes

The support vectors are the subset of data points that determine the max-margin separator and decide which side of the separator a test case is on. The Lagrangian multipliers corresponding to the support vectors are non-zero. A decision tree uses a tree-like graph to model decisions and outcomes, and can be used to determine worst, best, and expected values for different scenarios. Random forest is an effective machine learning algorithm that uses the idea of bagging.

Uploaded by

Murthi Mk
Copyright
© © All Rights Reserved
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
76 views

MLT Notes

The support vectors are the subset of data points that determine the max-margin separator and decide which side of the separator a test case is on. The Lagrangian multipliers corresponding to the support vectors are non-zero. A decision tree uses a tree-like graph to model decisions and outcomes, and can be used to determine worst, best, and expected values for different scenarios. Random forest is an effective machine learning algorithm that uses the idea of bagging.

Uploaded by

Murthi Mk
Copyright
© © All Rights Reserved
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
You are on page 1/ 2

SUPPORT VECTORS

The support vectors are used to decide which side of the separator a test case is on.
The support vectors are the subset of datapoints that determines the max-margin separator.
The Lagrangian multipliers corresponding to the support vectors are non-zero
The max-margin separator is not a non-linear combination of the support vectors.
For two dimensional data points, the separating hyperplane learnt by a linear SVM will be a straight
line
In regression the output is Continuous.
DECISION TREE
 A Decision tree is a decision support tool that uses a tree-like graph or model of decisions and their
possible consequences, including chance event outcomes, resource costs, and utility
 Decision tree is an example of linear classifier.
 Decision Tree algorithm is not an example of an ensemble method
 Entropy is a measure of purity
 An attribute with lower mutual information should be preferred to other attributes
 The entropy of a node typically increasesas we go down a decision tree.
Decision trees are prone to be overfit
Which of the following are the advantage/s of Decision Trees?
 Use a white box model, If given result is provided by a model
 Worst, best and expected values can be determined for different scenarios
 Possible Scenarios can be added

Random Forest  used and effective machine learning algorithm based on the idea of bagging?
The following property of a within-class scatter matrix is Non-singular must for LDA.

STATISTICS
o The expected value or Mean of a random variable is the center of its distribution.
 The measure of spread of a random variable is a varaiance .
 The square root of the variance is called the Standard deviation.
 Kohonen’s learning ?it is type of learning in which, the feature maps, the weights are
updated for winning unit and its neighbour,

Naive Bayes :
 In Bayes Theorem, Class conditional probability is called as Likelihood
 Assumes that all the features in a dataset are equally important and that they are independent
 PCA is an example of Unsupervised Classification
 PCA is used for Dimensionality Reduction
 SVM
 Consider a point that is correctly classified and distant from the decision boundary. SVM will be
unaffected by this point

k-Nearest Neighbor
 k-NN algorithm can be used for classification and regression.

 k-Nearest Neighbor is a Non parametric, lazy algorithm

 k-NN performs much better if all of the data have the same scale
 k-NN works well with a small number of input variables (p), but struggles when the number of inputs is
very large

 k-NN makes no assumptions about the functional form of the problem being solved

LDA

 Linear Discriminant Analysis is Supervised Learning

 Gaussian function is also called Bell function.


 The span of the Gaussian curve is determined by the varianceof the distribution
 In Supervised learning, class labels of the training samples are known
 Gaussian mixtures are also known as Linear super-position of Gaussians
PCA:

 Ensure variables are independent of one another

 Reduce the number of variables,but are not able to identify variables to completely remove from
consideration

 Comfortable making your independent variables less interpretable

 The tool used to obtain a PCA is SVD

You might also like