EXISTING OUTLIER VALUES IN FINANCIAL DATA VIA WAVELET TRANSFORM

Outlier detection is one of the major problems of large datasets. Outliers have been detected using several methods such as the use of asymmetric winsorized mean. Al-Khazaleh et al. (2015) has proposed new methods of detecting the outlier values. This is achieved by combining the asymmetric winsorized mean with the famous spectral analysis function which is the Wavelet Transform (WT). Thus, this method is regarded as MTAWM. In this article, we will expand this work using the modern Wavelet function known as the Maximum Overlapping Wavelet Transform (MODWT). The results of the study shows that after comparing the new technique with the previous mentioned techniques using financial data from Amman Stock Exchange (ASE), the Maximum overlapping wavelet transform- asymmetric winsorized mean (MWAW) was considered the best method in outlier detections