Data Mining

Data Mining Methodologies



Table of Content

INTRODUCTION3

Classical methods: Statistics, Neighborhoods and Clustering3

Statistics4

Data, Counting And Probability4

Histograms5

Table I6

Figure I7

Figure 27

Statistics for Prediction9

Linear regression9

Figure 310

Nearest Neighbor12

Figure 413

Clustering13

Clustering for Clarity13

Table 214

Table 315

Table 416

REFERENCES17

Data Mining Methodologies

Introduction

            This overview presents a recount of some of the most widespread data mining algorithms in use today.   We have broken the consideration into two parts, each with a exact theme:

Classical methods: Statistics, Neighborhoods and Clustering

Next Generation Techniques: Trees, Networks and Rules (Witten 2000)

          Each part will recount a number of data mining algorithms at a high grade, focusing on the "big picture" in order that the book reader will be adept to realize how each algorithm aligns into the countryside of data mining techniques.   Overall, six very broad categories of data mining algorithms are covered.  Although there are several other algorithms and numerous variations of the methods recounted, one of the algorithms from this assembly of six is nearly habitually utilised in genuine world deployments of data mining systems. (Westphal 1998)

 Data mining involves the use of sophisticated data analysis tools to discover valid

patterns and connections in large facts and figures sets. It consist of more that collecting and managing data, it also consist of analysis and prediction. Data mining can be performed on data represented in quantitative, textual, and multimedia forms.

Classical methods: Statistics, Neighborhoods and Clustering

I took air journey the Boston to Newark shuttle lately and sat besides a lecturer from one the Boston locality Universities.  He was going to talk about the  drosophila (fruit flies) genetic makeup to a pharmaceutical business in New Jersey.  He had  amassed the world's biggest database on the genetic makeup of the crop proceed by plane and had made it accessible to other investigators on the internet through Java submissions accessing a bigger relational database.

The major techniques that we will discuss here are the ones that are utilised 99.9% of the time on existing enterprise problems.  There are absolutely many other ones as well as proprietary techniques from specific vendors - but in general the industry is converging to those methods that work consistently and are understandable and explainable.

Statistics

By firm delineation "statistics" or statistical methods are not data mining.  They were being utilised long before the period data mining was coined to request to enterprise applications.  However, statistical methods are propelled by the data and are utilised to find out patterns and construct predictive models.  From the users viewpoint, you will be faced with a attentive alternative when explaining a "data mining" difficulty as to if you desire to strike it with statistical procedures or other data mining techniques.  For this cause, it is significant to have some concept of how statistical methods work and how they can be applied. I took air journey the Boston to Newark shuttle lately and sat besides a lecturer from one the Boston locality Universities.  He was going to talk about the  drosophila (fruit flies) genetic makeup to a pharmaceutical business in New Jersey.  He had  amassed the world's biggest database on the genetic makeup ...