3KE3 CHAPTER 4///

Why data mining?

\-More intense competition at the global scale

\-Recognition of the value in data sources

\-Availability of quality data on customers, vendors, transactions, Web, etc.

\-Consolidation and integration of data repositories into data warehouses

\-The exponential increase in data processing and storage capabilities; and decrease in cost

\-Movement toward conversion of information resources into nonphysical form

What is data mining?

The nontrivial process of identifying valid, novel, potentially useful, and ultimately understandable patterns in data stored in structured databases.

What are the four types of patterns?

association, prediction, cluster (segmentation), and sequential (or time series) relationships

What are the three most common data mining processes?

1) CRISP-DM

2) SEMMA

3) KDD (Knowledge Discovery in Databases)

What are the six steps in the CRISP-DM data mining process?

1) Business understanding

2) Data understanding

3) Data preparation

4) Model building

5) Testing and evaluation

6) Deployment

What are the 5 steps in the SEMMA data mining process?

Sample, Explore, Modify, Model, Assess

What are the steps involved in KDD?

1) Data selection

2) Data cleaning

3) Data transformation

4) Data mining

5) Internalization

Provide examples of commercial data mining software tools

IBM SPSS Modeler, SAS Enterprise Miner, Statistica

Provide examples of free and/or open source software tools

KNIME, RapidMiner, R

What are the major characteristics and objectives of data mining?

\-Data is presented in a variety of formats

\-Data environment is usually client/server architecture or web-based IS architecture

\-the miner is often the end-user

\-DM tools combined with spreadsheets & other software development tools

\-parallel-processing used

What are associations?

commonly co-occurring groupings of things

What are predictions?

tell the nature of future occurrences of certain events based on what's happened in the past; experience and opinion-based; associated with forecasting

What are clusters?

Identify natural groupings of things based on known characteristics such as assigning customers in different segments based on their demographics and past purchase behaviors

What are sequential relationships?

discover time-ordered events

What are the three main categories of Data Mining?

prediction, association, and segmentation (clustering)

What is classification?

The objective is to analyze the historical data stored in a database & automatically generate a model that can predict future behaviour

What are the three main types of prediction

classification, regression, time-series

What are decision trees?

One data mining methodology is decision trees, which generate rules and classify data sets; a hierarchy of if/then statements

What is clustering?

the tendency to remember similar or related items in groups

What are two commonly used derivatives in association mining?

1) Link Analysis -> the linkage among many objects of interest is discovered automatically

2) Sequence mining -> relationships are examined in terms of their order of occurrence to identify associations over time

What is the difference between statistics and data mining?

Statistics - collects sample data to test the hypothesis

DM & Analytics - use all existing data to discover new patterns & relationships

What is the difference between CRISP-DM and SEMMA?

CRISP-DM: takes a more comprehensive approach; including understanding of the business & relevant data to DM projects

\
SEMMA: implicitly assumes that the DM project's goals and objectives and data sources have been identified and understood

Describe Knowledge Discovery in Databases (KDD)

process of using DM methods to find useful info and patterns in the data; in relation to data mining: DM involves using algorithms to identify patterns in data derived from the KDD process

What's the difference between classification and clustering?

Classification learns the function between the characteristics of things and their membership through a supervised learning process -both variable types presented to the algorithm

\
Clustering learns through an unsupervised learning process where only the input variables are presented to the algorithm

What are the factors considered in model assessment?

predictive accuracy, speed, robustness, scalability, interpretability

What is a simple split?

partitions data into two mutually exclusive subsets called a training set & a testing set

What is the main criticism of the simple split?

that it makes the assumption that the data in the two subsets are of the same kind

What is k-fold cross-validation?

aka rotation estimation; the complete data set is randomly split into 'k' mutually exclusive subsets of approximately equal size

What are some additional classification assessment methodologies?

1) Leave-one-out

2) Bootstrapping

3) Jackknifing

4) Area under the ROC curve