Objectives

The master course « Algorithms for Business Intelligence and Digital Marketing » will focus on important algorithms for BI. We will work in the morning on methods and algorithms and in the afternoon by groups of 3 (or 4) students on practical cases with real company data. This course is available for the students in BA.

We will go beyond the usual « black box » presentation of these modern methods. Indeed, it is crucial to have a global vision of these methods to make, for example, the link between the Marketing division and the IT division of a company. The goal will consist of having the minimal vocabulary and the conceptual understanding to be able to talk of Marketing Algorithms with data scientists and computer scientists. Thus, we will study carefully various classes of algorithms and understand the mathematical concepts behind them and explain how to use these algorithms in BI and Massive Data context.

For the practical cases, the student will be able to: make a literature review, propose and implement business analytics solutions, write a report and present it in a context of work in group.

Contents

We will learn how to construct efficient algorithms for the following topics:

• Graph algorithms and optimization;
• Recommendation systems for digital marketing;
• Clustering methods and data visualization;
• Dimension reduction and mapping;
• Data mining and editing distance;
• Approximation algorithms for the travelling salesman problem;
• Randomized algorithms and parallel algorithms.

For the practical cases, the students, organized in groups, will receive a problem from a "sponsor" of the project: company, research institute, etc. The job is to understand the problem, do the necessary research and develop one or more solutions. If necessary, the group will also have to organize working meetings with the sponsor.

The sponsors provide the data, research questions and / or possibly help in supervising the work of the group of students. Students may be required to sign confidentiality clauses. The student cannot register for the course without accepting these clauses.

This is not an internship and, unless otherwise agreed, it is not intended that the group of students work directly with the sponsor. Group work remains within the University of Lausanne.

References

Cormen, T. H., Leiserson, C. E., Rivest, R. L., & Stein, C. (2009). Introduction to algorithms. MIT press.

Leskovec, J., Rajaraman, A., & Ullman, J. D. (2014). Mining of massive datasets. Cambridge university press.

Murphy, K. P. (2012). Machine learning: a probabilistic perspective. MIT press.

Vercellis, C. (2009). Business intelligence: data mining and optimization for decision making. New York: Wiley.

Evaluation

First attempt

Exam:

Without exam (cf. terms)  

Evaluation:

At the end of the course, students will have to provide a detailed written report and give a presentation. These two works form the final grade of the group.

CA graded: continuous assessment, final grade according to the following weighting system: 80% for the report and 20% for the presentation.

Retake

Exam:

Without exam (cf. terms)  

Evaluation:

For the student, the retake consists in writing a complement to the final report, and present this complement.

Individual grade = Grade of the amended report