Download Data Mining by Johan Johansson

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts
no text concepts found
Transcript
Data Mining
By: Johan Johansson
Mining Techniques





Association Rules
Decision Trees
Clustering
Nearest Neighbor
Neural Networks
Association Rules

If the confidence for Bread -> Milk is 80% then
it does NOT imply the Milk -> Bread has a
confidence of 80%.

A good rule has both high confidence and high
support.
Decision Trees

Produces rules that are mutually exclusive as
opposed to Association rules where there might
be a lot of redundancy.

Rule induction goes from the bottom up and
collect all possible patterns that are interesting.

Decision trees do a greedy search, looking for
the best possible split on the next step.
K-Nearest Neighbors

Instead of selecting the nearest neighbor, let’s
take a “vote” from the K nearest neighbors.

Larger values of K reduce the effect of noise on
the classification, but make boundaries between
classes less distinct.
K-Nearest Neighbors
Main differences between Clustering
and Nearest Neighbor

Clustering is mostly used for consolidation while
Nearest Neighbor is also used for predictions.

Nearest Neighbor uses supervised learning,
which is defined by the problem so it can make
accurate and relevant predictions.
Neural Networks

Mimic the structure and functioning of the
human brain.

Using sophisticated pattern detection and
machine learning algorithms.

Hard to use and hard to implement but very
effective.
Simple Neural Network
Simple Neural Network (cont’d)

Age and income (nodes) is based on fuzzy logic,
where is the values are between 0 and 1,
inclusive.

The weights (links) determines the output,
which is between 0 and 1, inclusive.

0.47(0.7) + 0.65(0.1) = 0.39 which is closer to 0
than 1 so it’s not the “default” in this case.
References


http://www.thearling.com/text/dmtechniques/
dmtechniques.htm
http://en.wikipedia.org/wiki/Data_mining
Related documents