General course objectives
Business Analytics (BA) is about exploring and analysing large
amounts of data to gain insight into past business performance in
order to guide future business planning. This course introduces a
portfolio of advanced data-centric methods which cover the three
main directions in BA: Descriptive (“what happened?”), predictive
(“what will happen?”), and prescriptive (“what should happen?”).
The methods will be applied to various business cases with aim to
demonstrate how to extract business value from data, provide
data-driven decision support along with effective data management
principles. In this course, advanced machine learning techniques
are used so understanding of data science and machine learning
basics is required as well as a good level of programming skills is
expected.
Learning objectives
A student who has met the objectives of the course will be able to:
- Identify business and societal impact opportunities related to
effective data utilization
- Summarize the identifying characteristics of advanced machine
learning approaches for descriptive, predictive and prescriptive
analytics
- Select and apply appropriate machine learning (regression,
classification, reinforcement learning, clustering) and data
management tools (e.g Pandas)
- Conduct small-scale machine learning experiments and understand
related scaling principles
- Apply one or more explainable AI techniques (e.g SHAP, Lime) in
data-driven decision advisory situations
- Conduct basic analysis of information in a natural language
form
- Understand the technical principles and potential of foundation
models (e.g. LLMs)
- Quantify uncertainties in predictive modelling, through
quantile regression and heteroskedasticity model
- Understand and be able to explain causal vs correlational
relationships between variables
- Be able to provide a clear and informative summary (executive
summary) for data-driven analyses and tools, including insights for
business and critical questions
Content
The classes are taught in an interactive manner, with theoretical
parts, intermingled with practical exercises. The practical
exercises are done in Python.
The main topics covered in the course includes web data mining;
natural language processing; recommender systems; explainable AI;
deep learning; reinforcement learning; spatio-temporal prediction
models; ensemble models; survival analysis; prediction uncertainty
and causality.
Course literature
Each module corresponds to its own recommended texts, that will be
provided in the end of the lecture notes. The mandatory texts shall
be publicly available online.
Last updated
02. maj, 2024