Scribd
Upload
10 views13 pages

Evaluating Data Mining Suitability

This document discusses applying data mining to problems and determining if data mining is suitable. It explores four types of knowledge - shallow, multidimensional, hidden, and deep - that can help determine if a problem is well-suited for data mining. The document also compares data mining to data querying and expert systems, providing an example of how data mining differs from simple querying. It outlines issues with nearest neighbor classification and discusses various algorithms that can be used for different data mining tasks.

Uploaded by

spotanand9941
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
10 views13 pages

Evaluating Data Mining Suitability

This document discusses applying data mining to problems and determining if data mining is suitable. It explores four types of knowledge - shallow, multidimensional, hidden, and deep - that can help determine if a problem is well-suited for data mining. The document also compares data mining to data querying and expert systems, providing an example of how data mining differs from simple querying. It outlines issues with nearest neighbor classification and discusses various algorithms that can be used for different data mining tasks.

Uploaded by

spotanand9941
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Applying Data Mining

By Susan L. Miertschin

Is Data Mining Appropriate for the Problem at H d? P bl t Hand?


Can you clearly define the problem? Does potentially meaningful data exist? Does the data contain hidden knowledge or is the data factual

and useful for reporting purposed only? d f lf d l? Will the cost of processing the data be less than the likely increase in profit seen by applying any potential knowledge gained from the data mining project?

Four Types of Knowledge can Help to Determine if the Problem is Suitable for Data Mining or Data Query
Shallow Knowledge Multidimensional Knowledge l d l l d Hidden Knowledge Deep Knowledge

Four Types of K F T f Knowledge l dg


Shallow Knowledge
Factual Easily stored and

Multidimensional Knowledge
Factual On-line analytical Processing

manipulated in a database
Hidden Knowledge
Patterns or regularities in data Not easily found using database query Data mining algorithms can find

(OLAP) tools used to manipulate multidimensional knowledge

Deep Knowledge
Knowledge stored in a database Can only be found if some

patterns

direction about what we are looking for is known

Data Mining vs. Data Query: An Example E l


Data Query
You already almost know

Data Mining
Find regularities in data Regularities are not

what you are looking for

obvious without the aid of tools


Wh ? Why? Amount of data Organization of data obscures patterns b Limits of human capabilities to consider many things at once

Expert Systems or Data Mining?


Expert System
A computer program that

Knowledge Engineer
A person trained to

emulates the problemproblem solving skills of one or more human experts

interact with an expert in order to capture the experts implicit knowledge in explicit form

Data

Data Mining Tool

If Swollen Glands = Yes Then Diagnosis = Strep Throat

Human Expert

Knowledge Engineer

Expert System Building Tool

If Swollen Glands = Yes Then Diagnosis = Strep Throat

Figure 1.2 Data mining vs. expert systems

What s Whats Wrong With Simple Search?


What is simple search? p
Nearest neighbor classifier

K-nearest neighbor classifier

Nearest Neighbor Classifier


Create a table of instances with known classifications
Th is the training d This h data

Get a new instance Compare the new instance to each instance in the training set

using the Euclidean distance metric for comparison purposes 2 2

x1 y1 xn yn

Find the instance in the training set that is closest on the

basis of the distance metric to the new instant Classify the new instance the same way as the one closest to it in the training data

Problems with Nearest Neighbor Classification Cl ifi ti


Computation times will be large when the training set is

large No differentiation of relevant from irrelevant attributes N way to tell which attributes d ff No ll h h b differentiate among classes l

10

Data Mining Algorithms


Different algorithms are available for different data mining tasks g g Different tools exist that implement different algorithms and different versions of algorithms

11

e.g., Algorithms Available in Microsofts Analysis Services A l i S i


Decision Trees Linear Regression Nave Bayes Clustering Algorithms Association Rules Sequence Clustering Time Series Analysis N Neural Networks lN k Logistic Regression
12

Applying Data Mining


By Susan L. Miertschin

13

You might also like