Machine Learning from Exploratory Data Analysis to Deep Learning

Which is just a quick recap of things I already know as it combines basic statistics like mean and correlation with hypothesis testing. This will be the basics of my data exploration for the dataset chosen above .

Using the chosen dataset, after having chosen a couple of variables of interest that emerge from the exploratory data analysis section, I will then employ linear regression analysis to understand the causality and relationships between these variables, to check and see if they are significant or not.

I will split the data into training and test sets to avoid overfitting, and employ regularization techniques such as ridge, LASSO and elastic net testing which one is more appropriate given the data.

I will explore different types of regressions from linear to polynomial to logistic.

After the regression section I will come up with questions based on the results and will want to use cluster analysis in order to see how different data groups interact. I will employ K-Nearest Neighbors, Support Vector Machines, decision trees and ensemble models.

I will further refine the clustering analysis by employing dimensionality reduction and principal component analysis

Optimize data with neural networks, from Recursive Neural Networks (RNNs) to Long-Short Term Memory Networks (LSTM) to Convolutional Neural Networks (CNNs).

Deep learning is a novel application of neural networks so it might not be needed, nonetheless I’ll use the data to build Generative Adversarial Networks (GANs) and I will use reinforcement learning by using rewards instead of using a method to minimize error.