Description: This project involves exploratory data analysis and predictive modeling using various statistical and machine learning techniques. In the financial domain, we analyze the Weekly dataset, containing weekly returns spanning two decades. We aim to identify patterns and trends in the data, perform logistic regression, and compare different classification algorithms like LDA, QDA, and KNN to find the most accurate method.
In the second part of the project, we shift our focus to the Auto dataset, which contains information about cars' attributes and gas mileage. Our goal is to predict whether a car gets high or low gas mileage based on these attributes. We create binary variables, visualize associations using graphs, split data into training and test sets, and apply LDA, QDA, and logistic regression for predictive modeling. Additionally, we explore the impact of K values in the KNN classifier on test accuracy.
This project provides an in-depth analysis of financial time series data and demonstrates the application of classification algorithms for predicting car mileage categories. It offers insights into which methods yield the best results and highlights the significance of feature selection and model evaluation in predictive modeling scenarios.