Description

Data mining is based on the mastery of fundamental data exploration techniques: descriptive, predictive or exploratory statistics. This practical course will introduce you to methods such as regressions and PCA and teach you how to implement them with R software.

Who is this training for ?

For whom ?

Infocentre / Datamining / Marketing / Quality managers, users and business database managers.

Prerequisites

Training objectives

Understand the principle of statistical modeling Choose regression type based on data type Making predictions Create selections and rankings in large volumes of data to identify trends

Training program

• Introduction to modeling
  • - Modeling: regression.
  • - Statistical modeling: reminders of statistical tests.
  • - Data analysis.
  • - Introduction to R software.
  • - Practical work Presentation of several modeling examples.
  • - Installation of R and the packages to be used.
  • - Applications on R, tests and interpretations on examples .
• Linear regression analysis
  • - Principle of linear regression.
  • - Simple regression, when the model has a single parameter for continuous data.
  • - Multiple regression, when there are more than 'a parameter.
  • - Other types of models for continuous data.
  • - Practical work Practical application in R.
  • - Case of simple regression and regression multiple.
• Logistic regression analysis
  • - Presentation of the different types of logistic regression.
  • - Binary logistic regression.
  • - Ordinal logistic regression.
  • - Multinomial logistic regression.
  • - Practical work Application on R with practical cases for cases of non-continuous data.
  • - Processing on data with two modalities, then with ordinal modalities, then nominal modalities.
• Component analysis
  • - Presentation of the different types of analyzes and selection.
  • - Principal Component Analysis (PCA).
  • - Multiple Correspondence Analysis (MCA).
  • - Hierarchical Classification on Principal Components (CHCP).
  • - Practical work The principal components make it possible to understand the covariance structure of the initial variables and/or to create a smaller number of variables to using this structure.
  • - Applications on R.
• Factor analysis of data
  • - Understand the principle of factor analysis: summarize the structure of data into a fewer number of dimensions.
  • - Factor Correspondence Analysis (CFA).
  • - Analysis Multiple Factor Analysis (AFM).
  • - Factor Analysis for Mixed Data (AFDM).
  • - Practical work Factor analysis exercises on R.
  • - Identification underlying "factors" of dimensions associated with significant variability.