Practical Approaches to Causal Relationship Exploration

2015
ISBN: 978-3-319-14432-0
Verlag: Springer International Publishing

Buch, Englisch, 80 Seiten, Format (B × H): 155 mm x 235 mm, Gewicht: 1533 g

Reihe: SpringerBriefs in Electrical and Computer Engineering

This brief presents four practical methods to effectively explore causal relationships, which are often used for explanation, prediction and decision making in medicine, epidemiology, biology, economics, physics and social sciences. The first two methods apply conditional independence tests for causal discovery. The last two methods employ association rule mining for efficient causal hypothesis generation, and a partial association test and retrospective cohort study for validating the hypotheses. All four methods are innovative and effective in identifying potential causal relationships around a given target, and each has its own strength and weakness. For each method, a software tool is provided along with examples demonstrating its use. Practical Approaches to Causal Relationship Exploration is designed for researchers and practitioners working in the areas of artificial intelligence, machine learning, data mining, and biomedical research. The material also benefits advanced students interested in causal relationship discovery.

Zielgruppe

Research

Autoren/Hrsg.

Li, Jiuyong

Le, Thuc Duy

Liu, Lin

Fachgebiete

• Mathematik | Informatik EDV | Informatik Daten / Datenbanken Data Mining
• Mathematik | Informatik EDV | Informatik Informatik Künstliche Intelligenz Maschinelles Lernen

Inhaltsverzeichnis

Introduction.- Local causal discovery with a simple PC algorithm.- A local causal discovery algorithm for high dimensional data.- Causal rule discovery with partial association test.- Causal rule discovery with cohort studies.- Experimental comparison and discussions.