Skip to Main content Skip to Navigation
Theses

Analyse Quantifiée de la Marche : extraction de connaissances à partir de données pour l'aide à l'interprétation clinique de la marche digitigrade

Abstract : Clinical Gait Analysis (CGA) is used to identify and quantify gait deviations from biomechanical data. Interpreting CGA, which provides the explanations for the identified gait deviations, is a complex task. Toe-walking is one of the most common gait deviations, and identifying its causes is difficult. This research had for objective to provide a support tool for interpreting toe-walker CGAs. To reach this objective, a Knowledge Discovery in Databases (KDD) method combining unsupervised and supervised machine learning is used to extract objectively intrinsic and discriminant knowledge from CGA data. The unsupervised learning (fuzzy c-means) allowed three toe-walking patterns to be identified from ankle kinematics extracted from a database of more than 2500 CGA (Institut Saint-Pierre, Palavas, 34). The supervised learning was employed to explain these three gait patterns through clinical measurement using induced rules from fuzzy decision trees. The most significant and interpretable rules (12) were selected to create a knowledge base that has been validated in terms of the literature and experts. These rules can be used to facilitate the interpretation of toe-walker CGA data. This research opens several prospective paths of investigation, ranging from the development of a generic method based on the proposed method for studying movement to the creation of a pathologic gait simulator.
Document type :
Theses
Complete list of metadata

https://tel.archives-ouvertes.fr/tel-00010618
Contributor : Stéphane Armand <>
Submitted on : Thursday, October 13, 2005 - 4:46:14 PM
Last modification on : Friday, March 26, 2021 - 10:59:27 AM
Long-term archiving on: : Tuesday, September 7, 2010 - 5:39:10 PM

Identifiers

  • HAL Id : tel-00010618, version 1

Collections

Citation

Stéphane Armand. Analyse Quantifiée de la Marche : extraction de connaissances à partir de données pour l'aide à l'interprétation clinique de la marche digitigrade. Sciences du Vivant [q-bio]. Université de Valenciennes et du Hainaut-Cambresis, 2005. Français. ⟨tel-00010618⟩

Share

Metrics

Record views

1194

Files downloads

19969