Hello Folks!
I have been going through various articles about fitting a Data Science / ML model to gain the highest accuracy possible. The various techniques to do model preprocessing as to delete the unwanted data, fill null values , getting the exact hyperparameters and so on.
But no article was significantly proposed to see how the person has reached the “Final” fitted model, why he has chosen only those specific algorithms and not others, simultaneously explaining the differences in choosing the algorithms. No article was written about how choosing different algorithms or as I can say rejection of some algorithms could affect the accuracy of the final fitted model.
What are the effects of these algorithms on different stages of a model building naming: preprocessing, feature selection and so on? They have just given us the techniques which finally went well for a specific dataset.
That’s why I am here to propose a series of 50-plus articles to do what these people have not done.
The salient features of my articles will be :
Datasets from different domains which could solve real-life problems.
A dataset will be passing through various paths and each path will have various algorithms in each model-building stage. This in turn will lead to various “Final Models” for a specific dataset.
Since various algorithms will be used for a specific stage of model building, I will be justifying with reasons why a specific algorithm should be used and should not be used. This, in turn, will lead us to simultaneously see what is the effect of these algorithms on the accuracy of the final model in a specific path.
I will also be providing a detailed whimsical wireframe of a specific dataset which can give us a brief idea as to what is done in a specific path. What all algorithms have used and what accuracies have we got by following a specific path?
The goals I will be achieving from it are :
To give the beginners an idea to develop reasoning behind the use and not use of a specific algorithm in a specific dataset.
To go through and understand the use of different algorithms to see the charm of machine learning.
This way I will be contributing to the Machine Learning community as much as possible.
Thanks for reading. Make sure to comment if you want to clarify anything. Make sure to follow me and share it with your friends for upcoming updates.Will see you into the real world of fitting datasets.