Outlier Detection Technique for Univariate Normal Datasets

This paper presents an outlier detection technique for univariate normal datasets. Outliers are observations that lips abnormal distance from the mean. Outlier is a useful in such areas as fraud detection, financial analysis, health monitoring and Statistical modelling. Many recent approaches detect outliers according to reasonable, pre-defined concepts of outlier. Methods Gaussian method have been widely used data-sets, however, methods use measure central tendency dispersion affected by hence making be less robust towards outliers. The study aimed at providing alternative can data sets deploying measures variation least (median geometric variation). formulated formula using median then applied formulation on randomly simulated dataset with recorded number detected comparison other two existing best detection. compared sensitivity three simulation was done different ways, first considered mean constant standard deviation while second test held varying deviation. performed best, eliminating most required techniques when there also established stricter varied but still stands out defined relative not more sensitive than Method well technique. In conclusion, could employed detections data-sets it almost same data-sets.