Advanced AI Techniques for Data Design and Analysis

Course category

Vertical

Level

Basic

Duration

24 hours [6 modules - 2 modules per day, 4 hours each, for 3 days]

Time

09:00 - 13:00 | 14.00 - 18.00

Language

English

Tutor

Vito Primavera

Description

The course “DOE Design & Data Analysis,” structured in 6 modules over 3 days, is aimed at technicians who wish to acquire high competence in both the design of experiments to be performed (in the laboratory or with numerical analyses) and the interpretation of the data obtained from them.

The first topic of the course, DOE Design (Design Of Experiments), focuses on learning the theoretical and practical knowledge necessary for generating experiments that maximize the quality of the information that the data can provide. This requires the application of algorithms capable of “smartly” positioning the experiments in the design space (also depending on the final purpose of the study), eliminating redundant observations on one hand and reducing the time/resources needed to perform the analyses/experiments on the other.

The second topic of the course, logically following the previous one, focuses on learning the theoretical and practical knowledge necessary for expert use of appropriate data analysis methodologies and tools. The goal is to manage the path “data → information → knowledge” which can result in the implementation of a more conscious and efficient optimization process or the knowledge itself useful for understanding one’s application.

The methodologies and tools employed fall within the realm of classical statistics, the Exploratory Data Analysis (EDA) approach, and Multi-Variate Analysis (MVA) techniques, each of which has its own usage prerogative but can also be used together to increase the level of confidence in the acquired knowledge.

The final topic covered focuses on acquiring technical and practical knowledge for the development of predictive models.

To make the course interactive through online modes, it is planned to carry out a series of hands-on activities immediately following the theoretical description of the topics covered, using commercial software.

Target Audience

The course is aimed at users (engineers of all specializations, designers, researchers) for whom it is relevant to:

• understand the behavior of the application of interest with the aim of managing its performance and reusing the acquired knowledge for similar applications.
• identify the “driving parameters” of their application, thereby increasing the efficiency of any optimization process or laboratory tests.
• make decisions based on objective criteria and justifiable results.

Pre-requisites

Basic knowledge of statistical analysis

Agenda

Module 1 – DOE Techniques (duration 4h) – Day 1/3

• DOE definition
• When and Why to apply DOE Techniques
• Main characteristics of DOE algorithms
• Application Examples

Module 2 – Statistical Analysis (duration 4h) – Day 1/3

• Importance of Statistical Analysis
• Correlation and Scatter Matrix
• T-Student Test
• Sensitivity Analysis (Pareto Plot)
• Application Examples

Module 3 – Statistical Analysis (duration 4h) – Day 2/3

• Linear and Interaction Effects
• Probability Density Function (PDF)
• Cumulative Distribution Function (CDF)
• Application Examples

Module 4 – Statistical Analysis (duration 4h) – Day 2/3

• Definition of Response Surface Model (RSM)
• Main Characteristics of RSM Algorithms
• Application Examples

Module 5 – Advanced Data Mining Techniques: MVA & PCA introduction (duration 4h) – Day 3/3

• Intro Multi Variate Analysis (MVA)
• Principal Component Analysis (PCA)
• Application Examples

Module 5 – Advanced Data Mining Techniques: Clustering & SOM introduction (duration 4h) – Day 3/3

• Clustering (H and partitive)
• Self-Organizing Maps (SOM)
• Application Examples