A clustering and visualization technique using Self Organizing Maps (SOM) has been developed to classify hits in High Throughput Screening (HTS) data. The cluster and cluster neighborhood information from the SOM results is used to identify what unique combination of structural features are similar in clusters of compounds that have high activity levels. The technique is also able to highlight anomalous compounds. An anomalous compound may show no activity despite close similarities to the high-activity compounds. Such anomalous compounds could be false negative results. Alternately, an anomalous compound may exhibit activity in the screen but is not structurally similar to other clusters of compounds. The SOM and visualization technique will be discussed in this presentation.