FONDAMENTI DI ANALISI DATI E LABORATORIO

Academic Year 2017/2018 - 1° Year - Curriculum Data Science

Teaching Staff

Credit Value: 9
Taught classes: 36 hours
Exercise: 24 hours
Laboratories: 12 hours
Term / Semester: 1°

Learning Objectives

Fundamental of Data Analysis
Basic introduction to state of the rt data analysis and automated classification.
Laboratory of Foundamental of Data Analysis
The objective of the course is the acquisition of knowledge of:
- Practical tools for the management and analysis of data;
- Tools for data visualization and exploration;
- Help the understanding of theoretical concepts and models through the implementation of algorithms and/or the analysis of existing implementations;
- Practical methodologies to train and use machine learning and data analysis algorithms to build automated systems for decision support;
- Tools to produce reports detailing a data analysis process.

Fundamental of Data Analysis
Data visualization, descriptive statistics

regression and correlation. Logistic regression.

Bayes approach to learning. MAP.

TS, CS, trainign and generalization error. Confuson matrix. ROC.

LDA, SVM.

Kernel trick: non linear SVM

PCA, non linear techniques for dimension reduction

K-nn

CART.

Clustering: k-means and hierarchical clustering.

Ensblem techniques. Boosting
Laboratory of Foundamental of Data Analysis
- Introduction to Python for scientific calculus;
- Descriptive statistics
- Gaussian distribution;
- Linear regression;
- Correlation and logistic regression;
- Probability distributions;
- Threshold based classification, ROC curves, confusion matrix;
- MAP classification, classification based on Mahalanobis distance, Naive Bayes classification
- Principal Component Analysis, Linear Discriminant Analysis
- Support Vector Machine
- K-Nearest Neighbour
- Clustering: K-Means and Mixtures of Gaussians
- Decision Trees for classification