This approach involves the statistical analysis of co-occurrence patterns of words in large corpora to derive their meaning. It focuses on the relationships between words and the contexts in which they appear, and has been successful in developing techniques for applications such as word sense disambiguation and sentiment analysis.