Similar Items: Unsupervised clustering of audio data for acoustic modelling in automatic speech recognition systems