7 - Indira Gandhi Centre for Atomic Research

8. Interpretation: Includes interpreting the discovered patterns and possibly returning
to any of the previous steps as well as possible visualization of the extracted
patterns, removing redundant or irrelevant patterns and translating the useful ones
into terms understandable by the users.
9. Using the discovered Knowledge: Includes incorporating discovered knowledge
into the performance system, taking action based on the knowledge or simply
documenting it for management/later use.
3. Knowledge Discovery Techniques
The Knowledge discovery can be of two categories: descriptive knowledge discovery and
predictive knowledge discovery. The former describes the data set in a concise and
summary manner and presents general properties of the data; whereas the later constructs
one or a set of models, performs inference on the available sets of data and attempts to
predict the behavior of new data sets. The features of knowledge discovery includes
• Large Amount of data
• Efficiency
• Accuracy
• Automated Learning
• High Level Language
• Interesting Results
3.1 Probabilistic Approach
This method utilizes graphical representation models to compare different representations.
Visualization tools are a class of advanced graphical presentation tools that facilitate data
exploration, hypothesis development, testing and evaluation of new or existing theory of
knowledge. Using appropriate computer technology, a skilled and well-motivated human
user can directly visualize patterns connection, correlations or lack thereof, among
parameters that have been measured or calculated. While this parameter can be codified in
numbers, visualization brings the enormous power of human perception to bear in entirely
new ways. One of the important uses of visualizing complex relationships among several
variables could be to understand the nature of empirical data in order to select the
appropriate techniques for further analysis.
3.2 Statistical Approach
The statistical approach uses rule discovery and is based on data relationships. An
inductive learning algorithm can be used to generalize patterns in the data and to construct
rules from the noted patterns. Online analytical processing (OLAP) is an example of a
statistically oriented approach.
3.3 Deviations and Trend Analysis
Pattern detection by filtering important trends is the basic for this approach. This is
normally applied to temporal databases, such as analysis of traffic on large
telecommunication networks.
143