Comprehensive Analysis and Forecasting of BMI, Grip Strength, and Sales Data

Software: R programming

Abstract

This report encompasses the application of statistical modeling and forecasting techniques on three distinct datasets, focusing on Body Mass Index (BMI) of Dutch boys, hand grip strength in English schoolchildren, and sales data from a UK-based company.

For the BMI dataset, various distributions were analysed, and the Exponentially Modified Gaussian (exGAUS) distribution was identified as the most suitable model. The hand grip strength data was best modeled using the Box-Cox t distribution (BCT), chosen after comparing multiple distribution models.

The sales data analysis involved selecting the best predictive model using the Generalised Akaike Information Criterion (GAIC), with the Box-Cox Cole and Green (BCCG) distribution emerging as the most effective. Predictions were made using a decision tree, achieving an accuracy score of 78.5%.

The report demonstrates the practical application of statistical distributions and machine learning models to real-world data, highlighting the process of data cleaning, model selection, and performance evaluation through various graphical representations and metrics. Peer review feedback suggested the inclusion of additional machine learning algorithms and more comprehensive graph interpretations.

Note: In case you need full R programming code text me at mpanocedric@gmail.com

Important link:

Open full pdf report