Lecture # 12 "Learning From Data"

The terminologies discussed in the chapter are: 
Learning: Learning is acquiring new, or modifying and reinforcing, existing knowledge, behaviors, skills, values, or preferences and may involve synthesizing different types of information. 
The Context of Learning:
The relevant technologies are:

·         Artificial Intelligence

·         Experts Systems

·         Case-Based Reasoning

·         Data Warehousing

·         Intelligent agents

·         Neural Networks

The Process Of Learning: It is a procedure which sort out and transforms data onto valid and practical knowledge.
The Goals Of Learning: The major goal of learning is to perk up the qualities of communiqué and decision making.
Learning From Data: There are two approaches to learn from data:
Top-down approach

·         Generate ideas

·         Develop models

·         Validate models

Bottom-up approach

·         Discover new (unknown) patterns

·         Find key relationships in data

Data Visualization: Data visualization is the study of the visual representation of data, information that has been abstracted in some schematic form, including attributes or variables for the units of information. According to Friedman (2008) "main goal of data visualization is to communicate information clearly and effectively through graphical means.
Artificial Neural Networks as Learning Model: The Artificial Neural Networks are modeled after human brain’s network. They simulate biological information processing via networks of interconnected neurons.
Learning In Neural Networks: There are two types of learning in NN:
Supervised: A teacher with set of examples of input and output.
Unsupervised: Does not need a teacher.
Business Applications: Following are the business applications of Neural Networks:

·         Risk Management

·         Predicting Foreign Exchange Fluctuations

·         Advance Evaluations

Association Rules: There are 4 types of association rules that help understanding the relationship that exist in data:

·         Boolean Rule: A rule that examines the presence or absence of items.

·         Quantitative Rule: A rule that considers the quantitative values of items.

·         Multi-dimensional Rule: A rule that refers to a multitude of dimensions.