Data Mining with Oracle 10g using Clustering and Classification Algorithms
Nhamo Mdzingwa · 2005
on which algorithm to use, in terms of which is the most effective and accurate algorithm in data mining, has always been a challenge for most data miners. The objective of this research is fundamentally focused on investigating the effectiveness of two algorithms available in Oracle10g for data mining. These are the K-Means and the O-Cluster algorithms. The second objective is to gather information from the dataset used in t he evaluation. Information gathering involves finding predictors of HIV AIDS prevention behaviour attributes. The results obtained are as follows; the first set is concerned with the evaluation of the K-Means and O-Cluster algorithms. Here it was observed that the O-Cluster algorithm builds more accurate models than the K-Means algorithm and also that the models by the O-Cluster algorithm find more accurate clusters when applied to new data. The second set o f results involves gathering information from the dataset. Here the attributes HIV Test and Know AIDS were identified as predictors of prevention behaviour of condom use and abstinence. These were found by distinguishing the clusters found in the dataset.