Machine Learning: 2 Books in 1: Machine Learning for Beginners, Machine Learning Mathematics. An Introduction Guide to Understand Data Science Through the Business Application

Data science strategies
Data science is mainly used in decision-making by making precise
predictions with the use of “predictive causal analytics”, “prescriptive
analytics” and machine learning.
Predictive causal analytics – The “predictive causal analytics” can be
applied to develop a model which can accurately predict and forecast the
likelihood of a particular event occurring in the future. For example,
financial institutions use predictive causal analytics-based tools to assess
the likelihood of a customer defaulting on their credit card payments, by
generating a model that can analyze the payment history of the customer
with all of their borrowing institutions.
Prescriptive analytics - The “prescriptive analytics” are widely used in the
development of “intelligent tools and applications” that are capable of
modifying and learning with dynamic parameters and make their own
“decisions”. The tool not only predicts the occurrence of a future event but
is also capable of providing recommendations on a variety of actions and its

---

resulting outcomes. For example, the self-driving cars gather driving-
related data with every driving experience and use it to train themselves to
make better driving and maneuvering decisions.
Machine learning to make predictions – To develop models that can
determine future trends based on the transactional data acquired by the
company, machine learning algorithms are a necessity. This is considered as
“supervised machine learning” which we will elaborate on later in this
book. For example, fraud detection systems use machine learning
algorithms on the historical data on fraudulent purchases to detect if a
transaction is fraudulent.
Machine learning for pattern discovery – To be able to develop models
that are capable of identifying hidden data patterns but lack required
parameters to make future predictions, the “unsupervised machine learning
algorithms”, such as “Clustering”, need to be employed. For example,
telecom companies often use the “clustering” technology to expand their
network by identifying network tower locations with optimal signal strength
in the targeted region.