Method and apparatus for creating a classifier indicative of a presence of a medical condition

An apparatus arranged to create a classifier indicative of a presence of a medical condition in a subject, wherein the apparatus is arranged to receive from a sensing unit chromatogram data indicative of a profile of volatile organic compounds in a sample from each of a first plurality of subjects having the medical condition and a second plurality of subjects without the medical condition, wherein the chromatogram data is indicative of a resistance of one or more metal oxide sensors associated with a gas chromatography column over a period of time;

• the apparatus comprising a processing unit and a memory unit storing computer executable instructions which, when executed by the processing unit, cause the processing unit to:
• select one of the chromatogram data as reference chromatogram data;
• align the remaining chromatogram data in relation to the reference chromatogram data;
• extract one or more features from the chromatogram data using a Mexican hat wavelet transform of a plurality of scales to determine a coefficient for the chromatogram data at each of the plurality of scales of the Mexican hat wavelet;
• selecting one of the plurality of scales as a best match for the chromatogram data based on an accuracy of a validation process;
• select one or more features of the chromatogram data indicative of the medical condition;
• and
• construct a classifier for determining a boundary between chromatogram data indicative of the medical condition and chromatogram data indicative of an absence of the medical condition.