Beruflich Dokumente
Kultur Dokumente
9.1 Introduction
The current business environment is constantly evolving. The global economic scenario is
providing opportunities as well as challenges. The factors affecting business environment
117
are consumer needs, globalization, and government policies, etc.
In such a business environment, organization basically has four action steps. The
organization can be reactive, anticipative, adaptive, or/and proactive. For this, organization
can develop a new strategy, get into partnership, etc.
Today most of the businesses are having a computerized business support. This support is
in form of decision support system, business analysis, etc.
The main objective of business intelligence is to bridge the gap between organization
current status and its desired position. Business intelligence helps organization achieve
commercial success along with sound financial management.
Business intelligence is framework designed to support decision-making process. This
framework combines architecture, database, analytical tools and applications. Business
analytics forms an integral part of business intelligence.
118
Business intelligence provides timely information for better customer relationship
management.
Business intelligence improves profitability of the company.
Business intelligence provides a facility of assessing organization‘s readiness in
meeting new business challenges.
Business intelligence supports usage of best practices and identifies every hidden
cost.
Business intelligence usage can be optimized by identifying key projects on which
company would like to focus. This process of highlighting key projects is called business
intelligence governance.
The importance of business intelligence is growing, and its usage has proliferated across
various types of users. Earlier, it was in the domain of IT staff, but now business team is
also independently handling business intelligence.
119
9.4 Business intelligence Tools
120
The next question which needs to be addressed is the source of the information. This
sourcing of data can be from enterprise itself, or it may be from the external sources. If the
source is within the organization than it is essential there is a common platform for all flow
of information.
The next question is the quantity of the data. Since there is large volume of data available,
based on the required company should gather data to have a normalized business behavior.
The next question is to make data valuable, once that is determined data management
comes from the picture, i.e. acquisition of data to retirement of data.
2. Transformation Tools
The required data needs to undergo ETL process. ETL process consists of extracting data,
transforming the data and loading the data. The process extracting data from the repository
is a straight-forward process. However, validation and cleansing of data is a difficult task.
This validation and cleansing of data is done through various well established business
rules. Transformation of data involves converting the data to standardized form.
3. Data Repositories
Organization can store data through data warehouses. Data warehouse sometime has data
mart, which is a partition to handle single business function. A metadata repository is used
to store data definition and technical information.
4. Analytical Tools and Presentation
There are several business tools available on the market, but it is essential to identify what
it intends to do with the data and then choose the tool.
5. Presentation Tools and Applications
Business intelligence can only work if end users are able to make sense out of that data.
Presentation tools should allow the users to manipulate complex data into to ad hoc reports
for company-wide distribution.
6. Operational Process
Operational Process determines how data management and business intelligence is to be
implemented within the organization. It deals with the question how the organization
creates manages data and different applications.
121
9.5 Data Mining
Data warehouses are large databases containing historical transactions and other data.
However, data warehouses in themselves are useless. To make data warehouses useful,
organizations must use software tools to process data from these huge databases into
meaningful information. Because executives can obtain significantly more information
about their customers, suppliers, and their own organizations, they like to call information
gleaned with such tools business intelligence (BI) or business analytics. The two main
uses of these databases are data mining and online analytical processing. These terms are
often used interchangeably by some people.
Data Mining
Data warehouses could be regarded as a type of mine, where the data is the ore, and new
useful information is the precious find. Data mining is the process of selecting, exploring,
and modeling large amounts of data to discover previously unknown relationships that can
support decision making. Data-mining software searches through large amounts of data for
meaningful patterns of information.
Data mining has four main objectives:
• Sequence or path analysis: Finding patterns where one event leads to another, later event.
• Classification: Finding whether certain facts fall into predefined groups.
• Clustering: Finding groups of related facts not previously known.
• Forecasting: Discovering patterns in data that can lead to reasonable predictions.
122
9.6 Data Mining Methods
Some data-mining tools are complex statistical analysis applications, and others use
additional tools which go beyond statistical analysis and hypothesis testing. While some
tools help find predefined relationships and ratios, they do not answer the question that
more powerful data-mining tools can answer: ―What are the relationships we do not yet
know?‖ This is because the investigator must determine which relationships the software
should look for in the first place.
Several types of analytical software are available: statistical, machine learning, and neural
networks.
Artificial neural networks: Non-linear predictive models that learn through
training and resemble biological neural networks in structure.
Genetic algorithms: Optimization techniques that use processes such as genetic
combination, mutation, and natural selection in a design based on the concepts of
natural evolution.
123
Decision trees: Tree-shaped structures that represent sets of decisions. These
decisions generate rules for the classification of a dataset. Specific decision tree
methods include Classification and Regression Trees (CART) and Chi Square
Automatic Interaction Detection (CHAID) . CART and CHAID are decision tree
techniques used for classification of a dataset. They provide a set of rules that you
can apply to a new (unclassified) dataset to predict which records will have a given
outcome. CART segments a dataset by creating 2-way splits while CHAID
segments using chi square tests to create multi-way splits. CART typically requires
less data preparation than CHAID.
Nearest neighbor method: A technique that classifies each record in a dataset
based on a combination of the classes of the k record(s) most similar to it in a
historical dataset (where k 1). Sometimes called the k-nearest neighbor technique.
Rule induction: The extraction of useful if-then rules from data based on statistical
significance.
Data visualization: The visual interpretation of complex relationships in
multidimensional data. Graphics tools are used to illustrate data relationships.
124