Improving Clustering on Occupational Text Data through Dimensionality Reduction