The purpose of this thesis is to explore the potential of the temporal features extracted from nasal airflow as the source for detecting inspiratory flow limitation non-invasively. This thesis focused on developing and validating algorithms capable of extracting physiologically meaningful features from nasal airflow recording, followed by classification of flow limited breaths, utilizing supervised and un-supervised learning. The feature values and the classification results were compared with the upper airway anatomical measurements such as neck circumference, upperairway cross-sectional area, and neck fluid volume. Three features (deviation index, peak amplitude variability, and peak number) were found to be significantly correlated with the upper airway measurements. Un-supervised classification suggested that participants with more neck fluid volume before sleep are associated with more flow limited breaths. The findings corroborated our understanding that inspiratory flow limitation is associated with upper airway anatomical abnormalities.