DOI: 10.36347/sajp.2012.v01i02.003

Pages: 74-80

Data from many fields rely on statistical analysis to help them make sense of it. There are two main schools of thought when it comes to statistical analysis: parametric and non-parametric. Although they share the goal of deducing conclusions from data, their assumptions and guiding principles are different. To help readers choose the right approach for any given situation, this article compares and contrasts the two approaches, including their advantages and disadvantages. The Parametric technique, which is also utilized in Machine Learning, is based on the assumption that a probability model may be determined using a set of defined parameters. When using a parametric technique, one must either have previous knowledge that the population is normally distributed or be able to quickly estimate it with a Normal Distribution, a feat made feasible by the Central Limit Theorem. One must be able to make assumptions about the data's population distribution in order to choose between non-parametric and parametric hypothesis testing. Non-Parametric analyses are more versatile (distribution-free), but they may not be as strong as parametric tests, which often need tighter assumptions.