Font Size:
OUTLIER DETECTION METHODS: A COMPARATIVE STUDY
Abstract
In data mining, data cleaning is the most important process for data analysis. Raw data that we collect from different sources may contain noise and missing values that may lead to erroneous results. Hence, data cleansing is required to clarify data from such anomalies. There are different techniques for handling missing values and identifying these outliers. In this paper different statistical techniques for outlier detection are studied and then implemented on weather dataset. A comparison is made at the end of paper to pick the best method. Weather datasets of Lahore, Pakistan was used for applying these methods and collected from online source.
Refbacks
- There are currently no refbacks.