Information Management course

News

2019-04-26: Results of the 2019-04-16 Exam.
2019-02-21: Results of the 2019-02-21 Exam.
2018-07-23: Results of the 2018-07-23 Exam.
2018-06-28: Results of the 2018-06-28 Exam.
2018-04-11: Results of the 2018-04-06 Exam.
2018-02-22: Results of the 2018-02-19 Exam.
2018-01-17: Results of the 2018-01-15 Written Exam
2017-12-29: Detailed information about exam has been added
2015-10-06: The 2015-16 version of the course has started
2014-10-02: The 2014-15 version of the course has started
2013-10-02: The 2013-14 version of the course has started
2012-10-05: Website is online

General Informations

Teacher:	A. Ceselli
Course schedule:	Tuesday 9.00 - 11.00 Room '3-nord'; Tuesday 11.00 - 13.00 Room '3-nord'
Tutoring:	after each lecture, or by appointment (send me an email)

Homepage of previous editions

The homepage of the 2012-13 edition of the course , homepage of the 2013-14 edition of the course , and the homepage of the 2014-15 edition of the course are still online.

Lectures support

Code summary of the 20/10/2017 R tutorial lecture
Dataset churn .
Dataset Breast Cancer Wisconsin (cat) .
Dataset Breast Cancer Wisconsin (diagnosis) .
Code summary of the 17/11/2017 FSA lab session

Lecture topics and slides (2016)

Disclaimer: the slides in many lectures are adapted from those of the book "Data Mining: concepts and techniques" by J. Han, M. Kamber and J. Pei (3rd edition) (see second slide of each file for proper reference).

Lecture 01 (Oct 6th '15): Introduction to Information Management and Analytics (slides in LibreOffice and (slides in pdf format); motivation and examples; big data and analytics; main issues in data mining.
Lecture 02 (Oct 7th '15): Information Management in Practice: Data Mining in Mobile Cellular Networks (M. Premoli).
Lecture 03 (Oct 13th '15): Know your data (slides in LibreOffice and slides in pdf format);
Lecture 04 (Oct 14th '15): R tutorials ( list of topics in pdf format , the R script and function definitions we developed during the lecture, with a few additional examples)
Lectures 05, 06, 07 and 08 (Oct 21th '15, Oct 23rd '15): Data preprocessing (slides in LibreOffice and slides in pdf format); dataset for exercises: Breast Cancer Wisconsin . An R IDE can be downloaded from the R Studio homepage .
Lecture 09 and 10 (Oct 29th '15): Data (dimensionality) reduction - Principal Component Analysis (slides in LibreOffice and slides in pdf format)
Lectures 11 and 12 (Nov 10th '15): PCA in practice; datasets for exercises: Houses dataset and Forest Fire dataset ; the R code of the PCA implementation we built in the lab session, and a few lines of script to test it.
Lectures 13 and 14 (Nov 17th '15): Data (dimensionality) reduction - Feature Selection Algorithms (slides in LibreOffice and slides in pdf format; datasets for experiments ).
Lectures 15 and 16 (Nov 24th '15): R programming and data (dimensionality) reduction - Feature Selection Algorithms. R code of the FSA implementation we built in the lab session.
Lectures 17 and 18 (Dec 1st '15): Mining frequent patterns (part 1: slides in LibreOffice and slides in pdf format; part 2: slides in LibreOffice and slides in pdf format).
Lectures 19 and 20 (Dec 10th '15): Classification Models and Methods ( General concepts, Decision Tree Induction: slides in LibreOffice and slides in pdf format; Bayesian Classifiers: slides in LibreOffice and slides in pdf format; Support Vector Machines, Evaluating Classification Models: slides in LibreOffice and slides in pdf format; ).
Lectures 21 and 22 (Dec 15th '15): Cluster analysis ( General concepts, Custering methods: slides in LibreOffice and slides in pdf format; Evaluating clustering results: slides in LibreOffice and slides in pdf format; ).
Lectures 23 and 24 (Dec 17th): Classification and Clustering in practice using R (R code of the lecture).
Uncovered during this edition of the course: Datawarehousing, Online Analytical Processing and Data Cube Technology overview (slides in LibreOffice and slides in pdf format).

Please, report me any error and typo you find in the slides.

Pointers to related material

The reference textbook for the course is "Data Mining: concepts and techniques" by J. Han, M. Kamber and J. Pei (3rd edition).
Lecture 8 is adapted from "Data Mining methods and models" by D.T. Larose .
Lectures 13 and 14 are adapted from L.C. Molina, L. Belanche, A. Nebot "Feature Selection Algorithms: a survey and experimental evaluation", proc. of IEEE ICDM 2002 and L. Belanche, F. Gonzales "Review and Evaluation of Feature Selection Algorithms in Synthetic Problems", arxiv 2011.
Lecture 20 is partially adapted from C. Vercellis "Business Intelligence: Data Mining and Optimization for Decision Making", Wiley and sons, 2011.
The former editions of the Information Management course were held by prof. A. Tettamanzi: you can find syllabus and slides at the corresponding website.

Details about the exam (starting from a.y. 2017/2018)

The course exam is composed by three parts: written theory, practical project and oral theory. Practical project and oral theory are optional.

Written Theory: it takes place in the day and time of the exam call. It is composed by very few open-ended questions , requiring some reasoning based on the theoretical part of the course. The result can be a score between 8 and 25, or FAIL.
Practical Project: it needs to be submitted by email to the teacher before the day of the exam call. It can be either an application chosen by the student and approved by the teacher, or an analysis of a dataset chosen from the UCI machine learning repository . The project needs to be described in a report. The format of the report is free, but as a suggestion the student can fill the following template in libreoffice format or template in word format . In writing the report focus more on your own interpretation of the results than on the technical way of getting them; rewriting the theoretical background for each step of the analysis is unnecessary. The result can be a score between 0 and 5, or FAIL.
Oral Theory: it takes place a few days after the day of the exam call. The students wishing to undertake it are asked to contact the teacher by email. The teacher publishes a time schedule after scoring the written exams. It covers all the topics of the course, that is both theory and practice. The result can be a score between 0 and 5, or FAIL

The final evaluation is the sum of scores obtained in the parts which are undertaken, or FAIL if a FAIL score is incurred in any part (that is, a single FAIL yields an overall FAIL). If the sum is below 18, the overall result is FAIL. If the sum is above 30, the overall result is 30. The teacher may assign honors when the overall result is 30, independently on the sum of scores (that is, obtaining 31 or more does not necessarily grant honors).
It goes without saying that the above are general guidelines to sketch the evaluation process, but (as in all university exams in Italy) the teacher has the final responsibility of choosing an appropriate score for the student, that might also differ from the sum of single scores.

Previous written exam questions:

Written theory of January 15th, 2018 .
Written theory of February 19th, 2018 .

Further guidelines:

Take into account that you have six exam calls during the year. It is not needed to give the exam at the first call: plan your exams schedule carefully.If you have temporary health problems you may simply consider to register for another call. If you have more serious problems requiring assistance, you need to contact our central office for disability and obtain proper certification.
Concerning the topics. There are no special topics granting to pass the exam: all the topics that I have covered during the lectures are equally important, and might be expected in questions during the exam.
Always consider that exams target at checking your full understanding of the course, including the skill of comparing, generalizing, synthesizing, finding weak and strong points of the methodologies you have seen during the lectures. Therefore, expect questions that do not directly translate in rewriting a few slides, but require instead your personal reasoning.
Aiming to just "pass the exam" is the best way of failing. Aim instead to top marks.
If you wish more informations about the exam, please attend the lectures: I discuss about it several times.

In general, take into account that the exam is not a sports competition, for which strict scoring rules would be needed, and the teacher is not a judge blindly applying these rules. There are no rankings with winners and losers. The teacher is ready to acknowledge and evaluate shades of gray. Just prepare at the best of your skills, and be confident that the exam is designed to provide a fair evaluation.

Information Management Course 2017-18