Health Informatics

The course will address the following topics in 3 teaching modules:

Course Introduction
Research problem classification
Data sources and data types
Compute data statistics
Data summarization
Graphical Programming Tool: Orange
Data Preprocessing
Supervised/Unsupervised Learning
General modelling concepts and workflow
Classification Trees
Logistic Regression
Model Interpretation
(Linear Regression, Clustering)



MODULE 1: DATA EXPLORATION AND DESCRIPTIVE ANALYSIS
The first part will be very strictly connected with the biostatistics course and will involve using excel for presenting patient data and making sta-tistical comparisons.
· Research problems categorization
· Structured, Semi-structured data source
Tables: Records and features
Data Types (nominal, ordinal, ranked, discrete, continuous)
· Exploratory Data Analysis
Statistics, query, summarization
Graphical representation
References for this part are (numbers refer to BIBLIOGRAPHY section):
· Spreadsheets: a few tips (Ariel platform)
· [4] Chapter 2,3
· [1] Chapter 1
· [3] Chapter 2,3

MODULE 2: DATA PREPROCESSING FOR PREDICTIVE MODELS
Graphical Programming Tool: Orange
· Installation of the tool and interface exploration
· Simple use cases
Data Preprocessing
· Outliers, Missing Values, Data Representation, Standardization, Discretization, Feature Engi-neering, (Unbalance data)
References for this module are:
· [4] Chapter 2, 4

MODULE 3: INTRODUCTION TO STATISTICAL LEARNING FOR PRE-DICTIVE MODELS
Supervised/Unsupervised Learning
· Classification, Regression, Clustering
Model Workflow
· Bias/Variance and overfitting, Holdout method
Supervised Learning
· Feature selection
· Classification: Classification Tree, Logistic Regression
· (Regression: Linear Regression)
(Unsupervised Learning)
· Dimensionality Reduction: PCA, Data embedding
· Clustering
References for this module are:
· [4] Chapter 19, 20
· [5] Chapter 2, 3.1-3.3, 4.1-4.3, 5, 8, (12)