Unformatted text preview: . Service providers can use this information to
better organize their web pages.
Trend Analysis Several large databases have time series data (records accumulated over time) with
timestamp. For example company's sales data over several years/months, a
customer's credit card transactions over a period of time, a the fluctuations in stock
prices over a period of time, are all time series data. Such data can be easily
analyzed find certain historical trends. The objective of trend analysis algorithms
for data mining is to discover the patterns and regularities in the data evolutions
along with the dimension of time. Such patterns have often been found to be very
useful in making decisions for larger benefits in the future.
A simple application of trend analysis algorithm may be to analyze the historical
trend of fluctuations in the stock prices of various companies, to predict future
behavior. Such analysis results can be effectively used by investors in the stock
market for making investment decisions for better profitability.
Data Mining Techniques
Data mining adopts several techniques drawn from many different areas for
implementing the various types of algorithms-discussed above. Some of the most
popular techniques used for data mining are:
2. Machine learning,
4. Neural networks,
5. Fuzzy sets, and
6. Visual exploration
They are briefly described below. Most data mining systems employ multiple of
these techniques to deal with different kinds of data, different data mining tasks,
and different application areas.
Statistical methods are very useful in data mining. Usually, statistical models are
built from a set of training data. An optimal model, based on a defined statistical
measure, is searched among the hypothesis space. Rules, patterns, and regularities
are then drawn from the model.
Some of the commonly used statistical methods used in data mining are:
1. Regression analysis. It maps a set of attributes of objects to an output variable....
View Full Document
This document was uploaded on 04/07/2014.
- Spring '14