Outlier Detection Using A New Hybrid Approach On Mixed Dataset

Data mining is a process of extracting hidden and useful information from the data. Outlier detection is a fundamental part of data mining and has huge attention from the research community recently. An outlier is data object that deviates from other observations. Detecting outliers has important applications in data cleaning as well as in the mining of abnormal points for fraud detection, stock market analysis, intrusion detection, marketing, network sensors. Most of the existing research efforts focus on numerical datasets which are not directly applicable on categorical dataset where there is little sense in ordering the data and calculating distances among data points. Furthermore, a number of the current outlier detection methods require quadratic time with respect to the dataset size and usually need multiple scans of the data; these features are undesirable when the datasets are large. This thesis focuses and evaluates, experimentally, an outlier detection approach that is geared towards categorical sets. In addition, this is a simple, scalable and efficient outlier detection algorithm that has the advantage of discovering outliers in categorical or numerical datasets by per