Line of Best Fit
Any enthusiastic/wanna be/aspiring to be Data Scientist’s first algorithm would be Linear Regression. The fundamental understanding of the algorithm is to find Line of best fit through scatter plot of data points.My life for past couple of months , if I try to visualise feels like a scatter plot and me trying to find my best fit line through those chain of events. This page is also a reflection of me trying to draft my journey via lines of best fit!
As any story would begin, mine too starts with “ A long time ago….” . But the difference is that the Protagonist here is Flawed. She lacks self-commitment but always wants to try new things. Hence, it results in picking up multiple things and then trail of incomplete stuffs resulting in those scattered points in her life. Its a repetitive cycle:
- Initial few days start with a boost up mind and body. Ready to conquer the world .
- Mid days are trying to dodge hurdles present in the task. Feeling like I understood all the sassy quotes related to what is life?
- End days are liking some thing else and promising this baby that I will be back! Alas never had been True.
One day she had enough. All those piled up stuff kept mocking her. Adding to it she could not recollect anything she had even read in those intermediate works. The urge to break free from this mess had her drowning. Struggling to breathe free she decided to give herself just one day to make some change in her routine. The night before she had came across “A small change, repeated every day, can change everything.” So she went ahead and changed her routine just by 15 mins and by the end of day it had done wonders.
Every day small changes were made, every day seemed like a struggle, everyday had set backs , every night points were analysed, learning rate was altered and next days best fit line was predicted. She still has not obtained the accuracy nor has eliminated the outliers. But slowly over the time, steadiness has been achieved in the trend.
If you search for Linear Regression, the internet is flooded with blogs, articles, code implementation, usecases , etc. Great pages can be found related to it on medium itself. When I decided to implement this algorithm , my first thought was how silly of me to do something that everyone has already done. But then I had two datasets — my life and the Kaggle dataset. I built a simple flask application to predict stocks for Amazon and Netflix and forecast them. Little did I know, that the very thought of what would be different from those that have already written about this, was the LEARNING. Learning the math intuition, different videos, blogs made my “ to be 4 lines of code “ to “Lets write a story about it”. Not only did I complete my end to end task but have achieved some thing far more than the accuracy of the model. It’s the satisfaction of completing it. It made me understand that what is considered as a basic topic is actually a Pandora’s box. The more you dig , the more you get. Let not those tech blogs stop you from exploring , let it be a means to do more.
This page is a reminder to ME to keep going. But for now I bid myself Adieu!
PS: https://github.com/adshre/End2EndProjects_LinearRegression/tree/master