Classifying imbalance childhood brain tumours through 1H-MRS metabolite profiles remains a challenging problem.
We presented an alternative oversampling method, semi-synthetic wavelet oversampling (SSWO).
Different from the classic Synthetic Minority Oversampling TEchnique (SMOTE) that oversamples the metabolite profiles, SSWO used the wavelet processed 1H-MRS as the oversampled 1H-MRS, followed by quantification and classification.
As the result, SSWO can provide dramatically better classification performance than non-oversampled or classic oversampled metabolite profiles.
An optimal balanced classification accuracy is achieved as 96% and 72% from 84% and 52% for the 1.5T and 3T cohorts of childhood brain tumours, respectively.