Examining the comments associated with YouTube postings of songs from the later decades of the 20th century can be fascinating. Many older people express how nostalgic the music might make them feel for that time in their lives, and how it evokes a desire to be young again. It is interesting to understand whether they reflect a social phenomenon only possible through modern technologies. The aim of this paper is to make an initial investigation. YouTube videos for Number 1 songs from the British charts since the 1960’s were identified. Their comments were extracted and labelled as being nostalgic or not. Two Machine learning techniques from the GATE tool were applied to the data for different feature sets to find which technique performed best at classifying nostalgia. The results show that, with cross-validation, the Decision Tree Classifier outperformed the Naïve Bayes. Additionally, it is shown that the feature set has an influence on the accuracy.