Arborescence des pages
ConfigureOutils de l'espace
Aller directement à la fin des métadonnées
Aller au début des métadonnées

CES7AH - CES9AE

Statistical data processing

 

ECTS Credits : 4

Duration : 36 hours

Semester : S7

Person(s) in charge :

Olivier DECK, Professeur,  olivier.deck@mines-nancy.univ-lorraine.fr, Judith Sausse, Professeur,  judith.sausse@mines-nancy.univ-lorraine.frThierry VERDEL, Professeur,  thierry.verdel@mines-nancy.univ-lorraine.fr

Keywords :

 Exploring and visualising data - statistical analysis - modeling and predicating

Prerequisites : First year statistics course or equivalent

Goal : Model and simulate real systems

Program and contents :

In the 70s and 80s, the development of computer technology enabled information storage which, in its most classical form, resembled tables of data, usually of great size. In many fields (geology, meteorology, medicine, economy, marketing, quality control, form recognition etc), through data analysis we were able to extract some of this information and digest it, principally to aid the decision process, or more generally to comprehend in some way the nature of phenomenon pertaining to the data


These forecasting methods especially enabled a prediction of the future developments of a phenomenon through a model founded on its past behaviour and the relative context. Since the 90s the digitisation of information has led to an accumulation of considerable masses of stored information in digital, amorphous and dynamic databases of public and private institutions, all kinds of data such as figures, text, images, sounds etc. Datamining is symbolic of the industrialisation of data analysis allowing a true exploitation of the gold mine of commercial information: “extracting precious minerals from the swamp of data”.

Abilities : 

Level

Description and operational vocabulary

Know 

the

The main methods we will see are :
- Multiple regression, analysis of variance and logistical regression
- Analysis and anticipation of time series
- Factorial analysis of correspondences
- Automatic classifications and discrimination using decision tree analysis
- Non linear neuronal methods for anticipation, discrimination and classification

The application of these techniques throws up a number of questions for the user, essentially:

  • What types of problems can be dealt with?
  • Which method is best? 
  • Which data should I select?
  • What sort of results will it produce?
  • What are their limitations?
  • How can they be applied?

JMP and R software will be used systematically throughout this course, both in the lectures and the practical lessons; focusing on real case studies in various sectors.

Understand 

Apply 

Analyse

Summarise

Assess

Evaluation :

  • Writtent test
  • Continuous assessment
  • Oral presentation
  • Project
  • Written test
  • Aucune étiquette