Abstract: The flaw detection will detect data objects that are inconsistent with normal dataset. In additional to normal data, there exist negative outliers in many applications and data will be imperfectly labeled. This paper represents an outlier detection approach to address data with imperfect data labels into learning. Our past approach works in two steps. In the first step, we develop a pseudo training dataset by computing possible values of each example based on its local action. We introduce kernel k-means clustering method and kernel Local Outlier Factor-based method to compute the likelihood values. In the next step, we introduce the obtained possible values and limited abnormal examples into SVDD-based learning to produce a more accurate classification for global outlier detection. The proposed system has three approaches. They are Naive Bayes approach, Logistic regression and . For classification of dataset we go for these two approaches which makes easy to find outliers.