1. The Pilot | Regression Towards Mean.
Have you ever stared at the leaves of a tree while they’re fluttering due to the wind? I find it very fascinating, as there is a certain pattern to it; and still we cannot determine how the leaf is going to flutter during the next possible iteration.
And that’s the beauty of nature and the chaos around us, it follows certain set of rules, but still doesn’t repeat itself.
In our seemingly random and chaotic lives, we often struggle at taking decisions when we are dealing with ambiguity and all the decisions that we make are in ambiguous surroundings; it’s just that we’re more aware of some situations than others, which in return makes us confident about some decisions than others.
For example: Why is trying new burger at McDonalds easier than and ordering at a Restaurant where you’re trying a new cuisine for the first time so difficult? And Don’t you think if you keep visiting this restaurant over and over again, it’s going to get easier ordering at this place?
The point is, As we get more exposure and become more aware of an environment we obtain confidence while making the decision. And this is exactly how AI teaches itself; AI is like a new born baby teaching itself to walk, it penalizes itself for every wrong it does and rewards itself for every right move, eventually achieving the goal of walking.
In this world of Machine Learning and Artificial Intelligence, the best algorithm that a human can ever train is their own gut feeling.
And here at The 5th Quadrant we are going to see at the reflections of our lives as data points and train our intuitions about dealing with ambiguities and eventually understanding Machine Learning. I am no expert at this but, I am excited to share this perspective with all of you and learn myself while stitching these blogs together. And together, over time , we are going to get accustomed to this environment and get better, one baby step at a time. So let’s get started.
Average, Mean: is the sum of all samples divided by the number of samples. All of us know this definition of mean; but what makes it special is that it tells us about the central measure of the data. It gives us one portable number that speaks what is the centre point that our data revolves around. But that brings us to a very important, what actually is data?
Everything. Heights of people? Likes on a Post? Number of Dogs you pet in a day? Your scores in Flappy Bird?
Speaking of Flappy Bird, I remember everyone playing that game for hours trying to get through those pipes and get a high score every time we played that game. Even though we always wanted to beat the high score, we couldn’t do it. Majority of out scores were revolving around the mean, and as we went away from the mean the frequency of those score reduced, which is why we rarely attained our high score or our least score. A distribution of this sort is called as normal distribution.
Once we attained our high score we felt that we had accomplished something great and we were going to break this high score in the next round? But what happened, we couldn’t, and we felt that we have underperformed, similarly every time we performed better than a low score we felt that we have performed way better than the previous round. But, in both cases after an extreme performance, we went back to performing close to our average. And this is what statistics calls as Regression Towards Mean. We tend to perform close to our means, and if we had an extreme performance chances are that the next score is going to lie close to the mean.
It is very important to comprehend this, as we are treating high scores and low scores as a metric to success. Given enough time, anomalies are going to occur: Remember the surprising victory of Leicester City in the Premier League? what happened the following year? It went back to 12th position and the year after that it attained the 9th position. Near. Average.
Getting back to the example of Flappy Bird, remember how we performed in that game after being exposed to it for hours after hours: our high scores increased, and even our average performance did. The increase in the average scores is a better parameter of success than boasting over an occasional high score or crying over a very low score.
This is an important lesson I learnt in high school when I couldn’t beat my highest pointer in the successive semester. High Scores are great motivations, but the average performance in the very next iteration shouldn’t bum you out because thats how it is supposed to be, and it is this average we should constantly try to improve upon. The advice to never give up and keep doing the thing works, only because we don’t let the highs or lows determine who we are, but with time and awareness about the situations we constantly improve not only our decision making but also our performances, thus becoming one of the best people in the normally distributed population in the skill we want to acquire.
Bad days and Good days will occur but most of our days are going to be average, and we have to strive hard to make them better. So next time we do extremely well at something or very poorly at something, let’s not let it get to our head because it is just a data point that lies at the very extreme which is bound to occur, and also it doesn’t really speak the truth.
As a quick exercise, in the comment section: list one example where in life you’ve observed normal distribution and where you’ve seen regression towards mean occur in your own life.
And here we are, at the end of the pilot. We have hit the runway after having a look at normal distribution, mean and regression towards mean and are ready for take off. In the next blog we are going to see about what is standard deviation, talk about vectors and matrices and get some more insights upon normal distribution.
Thank you for reading the blog. Hope you liked it. Tell me how you liked it and And if there is any feedback you can let me know here, or Direct Message me on Instagram, or you can connect with me on Linked In. Thank you for making it till here. See you. Until next time.
