Data is as Intricate as it is Significant in the Modern World (Article Critique Sample)

Data is as Intricate as it is Significant in the Modern World

Data is as intricate as it is significant in the modern world. Exploiting this data effectively for optimal value requires data scientists to have a deep understanding of statistical theory. In the article “The 10 statistical techniques data scientists need to master”, Le (2017) describes the current fundamental statistical methods that data scientists require to grasp to enhance productivity. Although this article has a rather complicated inundation of statistics jargon, it is built around instructive statistical technique explanations, practical examples, and the merits and demerits of some statistical methods.  
Le (2017) acknowledges that data is most valuable under application the appropriate contextualization, analysis and organization. As such, the author admits that data scientists competent in statistical learning are pivotal to impactful data manipulation for industrial, scientific and financial ends. Before diving into the ten statistical methods, the author enumerates the difference between machine learning and statistical learning to curtail any misunderstanding. Le (2017) first defines the three most basic statistical techniques which require fulfilment of ordinary least squares criterion to suitably deploy linear models: linear regression, classification, and resampling methods. Apart from the definitions, the author describes the specific implementations in each method. Next, the author outlines alternative least squares methods that guarantee better model interpretability and prediction accuracy for linear models: subset selection, shrinkage and dimension reduction techniques. Le (2017) then defines non-linear models which fit data through successive approximations. Following non-linear models is an outline of tree-based methods which solve both statistical classification and regression problems using decision-trees. Le (2017) then proceeds to elucidate support vector machines (SVM), part of supervised machine learning models, that help solve statistical classification problems. Lastly, the author describes unsupervised learning techniques which cluster uncategorized data.