Blog

Error and Sufferings in Stats and Life model

The crossroads of analytics and philosophy provide useful insights into our world and help us make it better. When I teach regression equations (e.g., Y = a + bx + e) in my classes, I usually do not specifically refer to e, since students would normally expect error to be the default. The main aim of an analyst is to reduce errors, losses, or cost functions. Where Y is the outcome (target) variable to be predicted. However, the presence of e is required for the model to be valid. While thinking about this equation, I began to consider a philosophical concept of suffering.

As much as any research must have errors in it, philosophical views indicate that suffering is prior to our existence. In statistics, the errors can be reducible or irreducible. Reducible errors: By choosing appropriate models, these factors can be reduced to minimize omitted-variable bias, model specification errors, and data inadequacy. Irreducible error is either caused by measurement error, random error, or some factors that cannot be observed. On the same note, some types of suffering are subject to mitigation by our own decisions, whereas others are our prerogative. In building a statistical model or in living life, one cannot do away with all the errors or misery. Straightforward models are easier to understand and can help avoid problems like overfitting. Similarly, a simple life can bring more peace than a lavish one; simplicity may not always be cherished in society. Similar to least-squares modelling, simplicity cannot be the most obvious option, yet it often provides both simplicity and robustness.

According to Osho, it is the hardest aspect of life to lead an ordinary one, and only extraordinary individuals can do so, such as Gautam Buddha and Jesus Christ. New statistical models are also continually developed to address the shortcomings of the previous ones, but they create new problems. Such a cycle without an end is analogous to the complicated problems of life. For example, plastic not only addresses product durability problems but also creates significant environmental issues. Elaborate challenges in life are similar to the statistical methods for dealing with outliers, such as winsorization and normalization. The day-to-day life is usually filled with routine suffering, which is enhanced by extraneous distractions like social media. Nevertheless, once the noise is removed from the data, which enhances the model's performance, it is easier to focus on what we can manage, and this gives satisfaction and clarity. What does it do to you as an analyst or a seeker? When you accept the art of being simple and concentrate on what is controllable, you will discover simplicity in the statistical analysis of life as well as the management of the complexities of life. The least in data modelling or personal lifestyle can result in more understandable findings and serenity.