EXPLORİNG THE EFFECTİVENESS OF DİFFERENT DATA CLEANİNG TECHNİQUES FOR İMPROVİNG DATA QUALİTY İN MACHİNE LEARNİNG

Abstract

Good quality data is an essential part for the purpose of reaching an accurate and trusted machine learning model , However the present gained datasets in the real world usually contains some serious issues like wrong values , missing data , outliers or data noises , which can lead to the problem of producing wrong machine learning algorithms . the research explore the effectiveness of different data cleaning techniques in improving data quality for machine learning works . the research compares and estimate the vary ways for data cleaning technics and their performance such as handling missing values, outlier detection and removal, data normalization, and feature scaling. Through comparing between different datasets and observing their behavior , the research analyses the effect of each technics in the datasets and the subsequent impact in the production in the machine learning model. The result of this research is going to contribute and assets data scientists in the process of making a better design when preparing datasets for a machine learning model . by dedicating the correct data cleaning technics , the world can improved the reliability and the consistency of a machine learning models which fundamentally will lead to the improvement of decision making in a different ranges