Global high dimension outlier algorithm for efficient clustering & outlier detection

Outliers are the set of observations which are different from the other observations. Outliers are generally treated as abnormality, anomaly, deviation etc. in any data. Outlier detection is one of the major task in various data mining application like pattern detection, anomaly detection etc. If outliers are not detected in data, it generates deformity in data and also provides incorrect results. There are various techniques presented by the researchers to detect outlier in data, but these techniques are not able to provide global solution for the outlier analysis in the process. Thus an enhanced outlier detection technique called GHOA (Global High Dimensional Outlier Detection Algorithm) is proposed. In that technique first dataset is divided into different categories then proposed technique is applied on that categorized data, which provides result for outlier detection. A result analysis for the proposed technique is presented in section IV which shows that proposed technique provides better result as compare to the other existing techniques.